Research Article On the Performance of the Measure for Diagnosing Multiple High Leverage Collinearity-Reducing Observations

Size: px
Start display at page:

Download "Research Article On the Performance of the Measure for Diagnosing Multiple High Leverage Collinearity-Reducing Observations"

Transcription

1 Hndaw Publshng Corporaton Mathematcal Problems n Engneerng Volume 212, Artcle ID 53167, 16 pages do:1.1155/212/53167 Research Artcle On the Performance of the Measure for Dagnosng Multple Hgh Leverage Collnearty-Reducng Observatons Arezoo Bagher 1 and Habshah Md 1, 2 1 Laboratory of Computatonal Statstcs and Operatons Research, Insttute for Mathematcal Research, Unverst Putra Malaysa, 434 Serdang, Selangor, Malaysa 2 Department of Mathematcs, Faculty of Scence, Unverst Putra Malaysa, 434 Serdang, Selangor, Malaysa Correspondence should be addressed to Habshah Md, habshahmd@gmal.com Receved 2 August 212; Revsed 9 December 212; Accepted 9 December 212 Academc Edtor: Stefano Lenc Copyrght q 212 A. Bagher and H. Md. Ths s an open access artcle dstrbuted under the Creatve Commons Attrbuton Lcense, whch permts unrestrcted use, dstrbuton, and reproducton n any medum, provded the orgnal work s properly cted. There s strong evdence ndcatng that the exstng measures whch are desgned to detect a sngle hgh leverage collnearty-reducng observaton are not effectve n the presence of multple hgh leverage collnearty-reducng observatons. In ths paper, we propose a cutoff pont for a newly developed hgh leverage collnearty-nfluental measure δ D and two exstng measures δ and l to dentfy hgh leverage collnearty-reducng observatons, the hgh leverage ponts whch hde multcollnearty n a data set. It s mportant to detect these observatons as they are responsble for the msleadng nferences about the fttng of the regresson model. The mert of our proposed measure and cutoff pont n detectng hgh leverage collnearty-reducng observatons s nvestgated by usng engneerng data and Monte Carlo smulatons. 1. Introducton Hgh leverage ponts are the observatons that fall far from the majorty of explanatory varables n the data set see 1 4. It s now evdent that hgh leverage pont s another prme source of multcollnearty; a near-lnear dependency of two or more explanatory varables 2. Had 5 ponted out that ths source of multcollnearty s a specal case of collnearty-nfluental observatons; the observatons whch mght nduce or dsrupt the multcollnearty pattern of a data. Hgh leverage ponts that nduce multcollnearty are referred as hgh leverage collnearty-enhancng observatons whle those that reduce multcollnearty n ther presence are called hgh leverage collnearty-reducng observatons

2 2 Mathematcal Problems n Engneerng 6 1. Collnearty-nfluental observatons are usually ponts wth hgh leverages, though all hgh leverage ponts are not necessarly collnearty-nfluental observatons 5. It s very mportant to detect collnearty-nfluental observatons because they are responsble for msleadng concluson about the fttng of a regresson model, whch gves wrong sgn problem of regresson coeffcents and produces large varances to the regresson estmates. Not many studes have been conducted n the lterature on collnearty-nfluental measures and we wll dscuss these methods n Secton 2. Nonetheless most of the exstng methods are not successful n the detecton of multple hgh leverage collnearty-nfluental observatons although ther performances are consdered good for the detecton of a sngle observaton. Moreover these measures do not have specfc cutoff ponts to ndcate the exstence of collnearty-nfluental observatons 1. These shortcomngs motvated us to propose a new detecton measure n such stuaton. Notably, the proposed measure s based on the Dagnostc Robust Generalzed Potental DRGP method developed by Habshah et al. 11 and wll be presented n Secton 3. Secton 4 exhbts the development of the collneartynfluental observatons that can be classfed as hgh leverage collnearty-enhancng or collnearty-reducng observatons. Bagher et al. 1 presented numercal examples and a smulaton study to propose a novel hgh leverage collnearty-nfluental measure and a cutoff pont for the detecton of hgh leverage collnearty-enhancng observatons. The authors also recommended cutoff ponts for collnearty-nfluental measures ntroduced by Had 5 and Sengupta and Bhmasankaram 12. It s also mportant to dentfy hgh leverage collnearty-reducng observatons. However, these observatons are more dffcult to dagnose because they hde the effect of multcollnearty n the classcal analyss. Followng Had 13, Imon 14, and Habshah et al. 11, nsecton 5, we propose a cutoff pont for Bagher s et al. 1, Had 5, and Sengupta and Bhmasankaram 12 s measures to dentfy hgh leverage collnearty-reducng observatons. A numercal example and smulaton study are performed n Sectons 6 and 7, respectvely, to evaluate the performance of our proposed measure δ D and compare ts performance wth Had 5 and Sengupta and Bhmasankaram 12 s measures δ and l. Concluson of the study wll be presented n Secton Collnearty-Influental Measures Let consder a multple lnear regresson model as follows: Y Xβ ε, 2.1 where Y s an n 1 vector of response or dependent varable, X s an n p matrx of predctors n >p, β s a p 1 vector of unknown fnte parameters to be estmated and ε s an n 1 vector of random errors. We let X j denote the jth column of the X matrx; therefore, X X 1,X 2,...,X p. Furthermore, multcollnearty s defned n terms of the lnear dependence of the columns of X. Belsley et al. 15 proposed the sngular-value decomposton of n p X matrx for dagnosng multcollnearty as follows: X UDV T, 2.2

3 Mathematcal Problems n Engneerng 3 where U s the n p matrx n whch the columns that are assocated wth the p nonzero egenvalue of X T X s n p, V the matrx of egenvectors of X T X s p p, U T U I, V T V I, andd s a p p dagonal matrx wth nonnegatve dagonal elements, k j, j 1, 2,...,p, whch s called sngular-values of X. Condton number of X matrx denoted as CN s another multcollnearty dagnostc measures whch s obtaned by frst computng the Condton CI of the X matrx and s defned as k j λ max λ j, j 1, 2,...,p, 2.3 where λ 1,λ 2,...,λ p are the sngular values of the X matrx. The CN corresponds to the largest values of k j. To make the condton ndces comparable from one data set to another, the ndependent varables should frst be scaled to have the same length. Scalng the ndependent varables prevents the egen analyss to be dependent on the varables unts of measurements. Belsley 16 stated that CN of X matrx between 1 to 3 ndcates moderate to strong multcollnearty, whle a value of more than 3 reflects severe multcollnearty. Had 5 noted that most collnearty-nfluental observatons are ponts wth hgh leverages, but not all hgh leverage ponts are collnearty-nfluental observatons. He defned a measure for the nfluence of the th row of X matrx on the condton ndex denoted as δ, δ k k, 1, 2,...,n, 2.4 k where k s computed by the egenvalue of X and when the th row of X matrx has been deleted. Due to the lack of symmetry of Had s measure, Sengupta and Bhmasankaram 12 proposed a collnearty-nfluental measure for each row of observatons, defned as l log ( k k ), 1, 2,...,n. 2.5 Unfortunately, they dd not propose practcal cutoff ponts for δ and l and only mentoned the condtons for collnearty-enhancng and collnearty-reducng observatons. To fll the gap, Bagher et al. 1 suggested a cutoff pont for δ and l for detectng collneartyenhancng observatons as cut CEO Medan θ 3MAD θ, 1, 2,...,n, 2.6 where cut CEO s the Collnearty-Influental Measure cutoff pont for the dentfcaton of collnearty-enhancng observatons whereby θ can be δ or l. θ cut CEO for θ < s an ndcator that the th observaton s a collnearty-enhancng observaton.

4 4 Mathematcal Problems n Engneerng 3. Dagnostcs Robust Generalsed Potental for Identfcaton of Hgh Leverage Ponts The th dagonal elements of the hat matrx, W X X T X 1 X T, s a tradtonally used measure for detectng hgh leverage ponts and s defned as w x T ( X T X) 1x, 1, 2,...,n. 3.1 Hoagln and Welsch 17 suggested twce-the-mean-rule 2 p 1 /n cutoff ponts for the hat matrx. Had 13 ponted out that the leverage dagnostcs may not be successful to dentfy hgh leverage ponts and ntroduced a sngle-case-deleted measure, known as potental, and s defned as p x T ( X T X ) 1x, 1, 2,...,n 3.2 or p w 1 w, 1, 2,...,n, 3.3 where X s the data matrx X wth the th row deleted. Imon 14 ponted that potentals may be very successful n the dentfcaton of a sngle hgh leverage pont, but they fal to dentfy multple hgh leverage ponts. To rectfy ths problem, Imon 14 proposed a group deleton verson of potentals GP, known as generalzed potentals. Pror to defnng the GP, Imon 14 parttoned the data nto a set of good cases remanng n the analyss and a set of bad cases deleted from the analyss whch were denoted as R and D. Nonetheless, Imon s measure has drawbacks whch are due to the neffcent procedure that he used for the determnaton of the ntal deleton set D. To overcome ths shortcomng, Habshah et al. 11 proposed the dagnostc robust generalzed potental DRGP where the suspected cases bad cases were dentfed by Robust Mahalanobs Dstance RMD, based on the Mnmum Volume Ellpsod MVE. Rousseeuw 18 defned RMD based on MVE as follows: RMD X T R X T C R X 1 X T R X for 1, 2,...,n, 3.4 where T R X and C R X are robust locatons and shape estmates of the MVE, respectvely. In the second step of DRGP MVE, the GPs are computed based on the set of D and R obtaned from RMD MVE. The low leverage ponts f any are put back nto the estmaton data set after nspectng the GP proposed by Imon 14 whch are defned as follows: p w D for D, w D 1 w D for R, 3.5

5 Mathematcal Problems n Engneerng 5 where w D X T XT R X R 1 X. He suggested the cutoff pont of p as p > Medan( p ) ( cmad p ), 3.6 where c can be taken as a constant value of 2 or 3. The DRGP MVE have been proven to be very effectve n the dentfcaton of multple hgh leverage ponts. 4. The New Proposed Hgh Leverage Collnearty-Influental Observatons Measures As already mentoned n the precedng secton, the man reason of developng a new measure of hgh leverage collnearty-nfluental measure s due to the fact that the commonly used measures faled to detect multple hgh leverage collnearty-nfluental observatons. In addton, not many papers related to ths measure have been publshed n the lteratures. It s mportant mentonng that the collnearty-nfluental measure whch were proposed by Had 5 and Sengupta and Bhmasankaram 12 are related to the Had s sngle-casedeleted leverage measure 13. Snce the robust generalzed potentals that was developed by Habshah et al. 11 was very successful n the dentfcaton of multple hgh leverage ponts compared to other wdely used methods, Bagher et al. 1 utlzed a smlar approach n developng multple Hgh Leverage Collnearty-Influental Measure HLCIM. The proposed measure s formulated based on Sengupta and Bhmasankaram 12 s measure wth slght modfcaton whereby almost smlar approach of DRGP MVE 11 was adapted. Hence t s referred as HLCIM DRGP and denoted as δ D. Ths new measure s defned as follows: δ D log log log ( k D k D ( k k ) ( k D k D ) ) f D, #{D} / 1, f #{D} 1, D, 1, 2,...,n, f R, 4.1 where D s the suspected group of multple hgh leverage collnearty-nfluental observatons dagnosed by DRGP MVE, p,#{d} s the number of elements n D group, and R s the remanng good observatons. As such, followng Habshah et al. 11 approach, three condtons should be consdered n defnng δ D. Bagher et al. 1 summarzed the algorthm of HLCIM DRGP n three steps as follows. Step 1. Calculate DRGP MVE, p,for 1, 2,...,n. Form D as a hgh leverage collneartynfluental suspected group whereby ts members consst of observatons whch correspond to p that exceed the medan p 3MAD p. Obvously the rest of the observatons belong to R, the remanng group.

6 6 Mathematcal Problems n Engneerng Step 2. Compute hgh leverage collnearty-nfluental values, δ D, as follows. If only a sngle member n the D group, the sze of R s n 1, andd, calculate log k /k where k ndcates the condton number of the X matrx wthout the th hgh leverage ponts. In ths way, δ D l. If more than one member n the D group, calculate log k D /k D where k D ndcates the condton number of the X matrx wthout the entre D group mnus the th hgh leverage ponts, where belongs to the suspected D group. For any observaton n the R group, compute log k D /k D where k D refers to the condton number of the X matrx wthout the entre group of D hgh leverage ponts plus the th addtonal observaton of the remanng group. Step 3. If any δ D values for 1, 2,...,n does not exceed the cutoff ponts n 2.6, put back the th observaton to the R group. Otherwse, D group s the hgh leverage collneartyenhancng observatons. Bagher et al. 1 only defned the cutoff pont for θ to ndcate hgh leverage collnearty-enhancng observatons and they dd not suggest cutoff pont for collneartyreducng observatons. The authors consdered θ to be hgh leverage collnearty-enhancng observatons f θ s less than the cutoff ponts; that s medan θ 3mad θ for θ <, where c s a chosen value 3 and θ may be δ D, δ or l. Snce hgh leverage collnearty-reducng observatons are also responsble for the msleadng nferental statements, t s very crucal to detect ther presence. In the followng secton, we propose a cutoff pont for dentfyng hgh leverage collnearty-reducng observatons. It s mportant mentonng that not all δ D whch exceed the cutoff pont are hgh leverage ponts. Ths s true for the stuaton when δ D exceeds the cutoff pont but belongs to the remanng group, R. In ths stuaton, the observaton s consdered as collneartynfluental observatons snce they are not hgh leverage ponts. 5. The New Proposed Cutoff Pont for HLCIM (DRGP) Had 5 and Sengupta and Bhmasankaram 12 mentoned that a large postve value of ther collnearty-nfluental measures, δ and l, respectvely, ndcates that the th observaton s a collnearty-reducng observaton. However, they dd not suggest any cutoff ponts to ndcate whch observatons are collnearty-enhancng and whch are collnearty-reducng. Bagher et al. 1 proposed a nonparametrc cutoff pont for hgh leverage collneartyenhancng observatons. Ther work has nspred us to nvestgate hgh leverage collneartyreducng observatons among the observatons that correspond to postve values of hgh leverage collnearty-nfluental measures. Fgure 1 presents the normal dstrbuton plot of θ. Based on ths fgure, any value that exceeds medan θ 3MAD θ can be utlzed as a cutoff pont for θ. Hence, we propose the followng cutoff pont: cut CRO Medan θ 3MAD θ, 5.1

7 Mathematcal Problems n Engneerng 7 Collnearty-enhancng observatons Collnearty-reducng observatons 3Mad (θ ) Medan (θ ) +3Mad (θ ) Fgure 1: Normal dstrbuton plot of hgh leverage collnearty-nfluental measure. where cut CRO s the Collnearty-Influental Measure cutoff pont for Collnearty- Reducng Observatons. θ can be δ D, δ or l. θ cut CRO for θ > s an ndcator that the th observaton s a collnearty-reducng observaton. 6. A Numercal Example A numercal example s presented to compare the performance of the newly proposed measure δ D wth the exstng measures δ and l. An engneerng data taken from Montgomery et al. 19 s used n ths study. It represents the relatonshp between thrust of a jet-turbne engne y and sx ndependent varables. The ndependent varables are prmary speed of rotaton X 1, secondary speed of rotaton X 2, fuel flow rate X 3, pressure X 4, exhaust temperature X 5, and ambent temperature at tme of test X 6. It s mportant mentonng that, the explanatory varables of ths data are scaled before analyss n order to prevent the condton number to be domnated by large measurement unts of some explanatory varables. Pror to analyss of ths data, the explanatory varables have been scaled followng Stewart s 2 scalng method as x j x j Xj, 1,...,p, j 1,...,n. 6.1 There are other alternatve scalng methods whch can be found n Montgomery et al. 1, Stewart 2, andhad 5. The matrx plot n Fgure 2 and the collnearty dagnostcs presented n Table 1 suggest that ths data set has severe multcollnearty problem CN Wewouldlketo dagnose whether hgh leverage ponts are the cause of ths problem. As such, t s necessary to detect the presence of hgh leverage ponts n ths data set. The ndex plot of DRGP MVE presented n Fgure 3 suggests that observatons 6 and 2 are hgh leverage ponts. By deletng these two observatons from the data set, CN ncreases to It seems that these two hgh leverages are collnearty-reducng observatons. Theeffect of these two hgh leverage ponts on collnearty pattern of the data s further nvestgated by applyng δ D, δ and l wth ther respectve new cutoff pont ntroduced n 5.1 for detectng hgh leverage collnearty-reducng observatons. Fgure 4 llustrates

8 8 Mathematcal Problems n Engneerng X 1 X 2 X 3 X 4 X 5 X 6 Fgure 2: Matrx plot of jet turbne engne data set. Table 1: Collnearty dagnostcs of jet turbne engne data set. Dagnostcs r r r r r 16.7 r Pearson correlaton coeffcent r r r 26.2 r r r r r r 56.3 VIF > Condton ndex of X matrx > the ndex plot of these measures. Accordng to ths plot, all these three measures have ndcated that observatons 6 and 2 as hgh leverage collnearty-reducng observatons. Nevertheless, besdes observatons 6 and 2, they detect a few more observatons as collnearty-reducng observatons. It s nterestng to note that none of the observatons are detected as hgh leverage collnearty-enhancng observatons or collnearty-enhancng observatons. It s worth mentonng that we do not have any nformaton about the source of the two exstng hgh leverage collnearty-reducng observatons cases 6 and 2. Therefore, we cannot control the magntude and the number of added hgh leverages ponts to the data n order to study the effectveness of our proposed measures. In ths respect, we have modfed ths data set n two dfferent patterns followng 7. Habshah et al. 7 ndcated that n the collnear data set, when hgh leverages exst n just one explanatory varable or n dfferent postons of two explanatory varables; these leverages wll be collnearty-reducng observatons. Thus, the frst pattern s when we replaced observatons 5, 6, 19, and 2 of X 2 wth a fxed large value of 5. The second pattern s created by replacng the large value of 5 to X 2 for observatons 5, 6 and observatons 19, 2 of X 3. The DRGP MVE ndex plot for Fgure 5 reveals that observatons 5, 6, 19, 2 are detected as hgh leverage ponts for modfed jet turbne engne data set.

9 Mathematcal Problems n Engneerng 9 DRGP (MVE) Fgure 3: DRGP MVE ndex plot of jet turbne engne data set. Fgures 6 and 7 present the ndex plot of δ D, δ and l for the frst and the second pattern of the modfed jet turbne engne data set. The results of δ D n these fgures agree reasonably well wth Bagher s et al. 1 fndngs that when hgh leverage ponts exst n just one explanatory varable frst pattern or n dfferent postons of two explanatory varables second pattern n collnear data sets, these observatons are referred as collnearty-reducng observatons. For both patterns, δ D correctly dentfed that observatons 5, 6, 19, and 2 are hgh leverage collnearty-reducng observatons. However, for the frst pattern, both δ and l are not successful n detectng all of observatons; 5, 6, 19, and 2 as hgh leverage collneartyreducng observatons. In the frst pattern, they only correctly detected observatons 19 and 2 as hgh leverage collnearty-reducng observatons. However, none of the added hgh leverage collnearty-reducng observatons can be detected by these two measures n the second pattern. It s mportant to note that for the frst and the second patterns, the values of δ and l for the observatons 5 and 6, and observaton 19, respectvely are becomng negatve. Ths ndcates that for both patterns, δ and l have wrongly ndcated these observatons as suspected hgh leverage collnearty-enhancng observatons. 7. Monte Carlo Smulaton Study In ths secton, we report a Monte Carlo smulaton study that s desgned to assess the performance of our new proposed measure δ D n detectng multple hgh leverage collnearty-reducng observatons and to compare ts performance wth two commonly used measures δ and l. Followng Lawrence and Arthur 21, smulated data sets wth three ndependent regressors were generated as follows: x j ( 1 ρ 2) z j ρz 4, 1,...,n; j 1,...,3, 7.1 where the z j, 1,...,n; j 1,...,3 are Unform, 1. The value of ρ 2 whch represents the correlaton between the two explanatory varables are chosen to be equal to.95. Ths amount of correlaton causes hgh multcollnearty between explanatory varables. Dfferent percentage of hgh leverage ponts are consdered n ths study. The level of hgh leverage ponts vared from α.1,.2,.3. Dfferent sample szes from n 2, 4, 6, 1, and

10 1 Mathematcal Problems n Engneerng δ (D) a δ l b 2 c Fgure 4: plot of collnearty-nfluental measures for orgnal jet turbne engne data set. 3 wth replcaton of 1, tmes were consdered. Followng the dea of Habshah et al. 7, twodfferent contamnaton patterns were created. In the frst pattern, 1 α percent observatons of one of the generated collnear explanatory varables were replaced by hgh leverages wth unequal weghts. In ths pattern the explanatory varable and the observaton whch needed to be replaced by hgh leverage pont were chosen randomly. The second pattern s created by replacng the frst 1 α/2 percent of one of the collnear explanatory varable and the last 1 α/2 percent of another collnear explanatory varable wth hgh

11 Mathematcal Problems n Engneerng 11 DRGP (MVE) a DRGP (MVE) b Fgure 5: plot of DRGP MVE for modfed jet turbne engne data set, a pattern1, b pattern2. leverages wth unequal weghts. The two ndependent varables are also randomly selected and the replacement of the hgh leverage pont to the observatons n dfferent postons of explanatory varables was also performed randomly. Followng Habshah et al. 11 and Bagher et al. 1, the hgh leverage values wth unequal weghts n these two patterns were generated such that the values correspondng to the frst hgh leverage pont are kept fxed at 1 and those of the successve values are created by multplyng the observatons ndex, by 1. The three dagnostc measures δ D, δ and l wth the proposed cutoff pont were ntroduced to 5.1 and were appled to each smulated data. The results based on the average values are presented n Table 2. The α and HLCIO n Table 2 ndcate, respectvely, the percentage and the number of added hgh leverage collnearty-reducng observatons to the smulated data sets. Furthermore, the number of hgh leverage ponts whch s detected by DRGP MVE s denoted as HL. It s nterestng to pont out that the percentage of the hgh leverage pont, p detected by DRGP MVE denotedashlntable 2 s more than the percentage of the added hgh leverage collnearty-reducng observatons to the smulated data sets, α. However, by ncreasng the sample sze and the percentage of added hgh leverage ponts to the smulated data, both percentages became exactly the same. The CN1 and the CN2 ndcate the condton number of X matrx wthout and wth hgh leverage collnearty-reducng observatons, respectvely. Moreover, Cut θ 1 and Cut θ 2 represent the number of hgh leverage collnearty-reducng observatons and the number of collnearty-reducng observatons whch have been detected by cutoff θ. Table 2 clearly shows the mert of our new proposed measure for hgh leverage collnearty-nfluental measure exhbted n 4.1. It can be observed that no other measures that were consdered n ths experment performed satsfactorly except for our proposed measure. The smulated data sets have been created collnearly whch produced large values of CN 1, condton number of smulated data sets wthout hgh leverage ponts CN 1 > 3. The added multple hgh leverage collnearty-nfluental observatons reduces multcollnearty among the smulated explanatory varables; ths reducton may result from the smaller values of CN 2 compared to CN 1. It s mportant mentonng that the reducton of the CN 2 values for the second pattern was much more sgnfcant compared to CN 2 for the frst pattern. We can conclude that the nfluence of the added hgh leverage ponts to dfferent

12 12 Mathematcal Problems n Engneerng δ (D) a.15.1 δ l b 19 2 c Fgure 6: plot of collnearty-nfluental measures for the frst modfed pattern of jet turbne engne data set. postons of two explanatory varables for changng the multcollnearty pattern of smulated data, s more sgnfcant compared to the added hgh leverage ponts to only one explanatory varable. The results of Table 2 for the frst pattern of smulated data sets ndcate that for small sample szes n 2 our proposed measure could not ndcate the exact amount of hgh leverage collnearty-reducng observatons. However, by ncreasng the sample sze and the percentage of added hgh leverage ponts to the smulated data sets, the measure s capable

13 Mathematcal Problems n Engneerng 13 Table 2: Collnearty-nfluental measures for smulated data sets. a Measures n 2 n 4 Pattern1 Pattern2 Pattern1 Pattern2 α HLCIO HL CN CN Cut δ D Cut δ D Cut δ Cut δ Cut l Cut l b Measures n 6 n 1 Pattern1 Pattern2 Pattern1 Pattern2 α HLCIO HL CN CN Cut δ D Cut δ D Cut δ Cut δ Cut l Cut l c Measures n 3 Pattern1 Pattern2 α HLCIO HL CN CN Cut δ D Cut δ D Cut δ Cut δ Cut l Cut l

14 14 Mathematcal Problems n Engneerng δ (D) a δ b l c Fgure 7: plot of collnearty-nfluental measures for the second modfed pattern of jet turbne engne data set. of detectng the exact amount of added hgh leverage collnearty-reducng observatons. It s evdent by lookng at the value of Cut δ D 1 s exactly the same as HLCIO. On the other hand, the other two collnearty-nfluental measures, δ and l, faled to ndcate the exact amount of hgh leverage collnearty-reducng observatons. It s worth notng that all of these three measures also detect some ponts as collnearty-reducng observatons see the Cut θ 2 n Table 2, where θ s δ D, δ or l. Smlar results wll be obtaned f pattern 1 can be drawn for the second pattern of the smulated data sets. Compared to the frst contamnaton pattern, t s clearly seen that δ and l almost completely faled to detect ether

15 Mathematcal Problems n Engneerng 15 hgh leverage collnearty-reducng observatons or collnearty-reducng observatons. Our proposed measure dd a credble job where t s successfully detect hgh leverage collneartyreducng observatons for both contamnated patterns. 8. Concluson The presence of hgh leverage ponts and multcollnearty are nevtable n real data sets and they have an unduly effects on the parameter estmaton of multple lnear regresson models. These leverage ponts may be hgh leverage collnearty-enhancng or hgh leverage collnearty-reducng observatons. It s crucal to detect these observatons n order to reduce the destructve effects of multcollnearty on regresson estmates whch lead to msleadng concluson. It s easer to dagnose the presence of hgh leverage ponts whch ncrease the collnearty among the explanatory varables compared to those whch reduce collnearty. In ths respect, t s very mportant to explore a suffcent measure wth an accurate cutoff pont for detectng hgh leverage collnearty-reducng observatons. In ths paper, we proposed a precse cutoff pont for a novel exstng measure to detect hgh leverage collneartyreducng observatons. By usng an engneerng data and a smulaton study, we confrmed that the wdely used measures faled to detect multple hgh leverage collnearty-reducng observatons. Furthermore, our proposed cutoff pont successfully detects multple hgh leverage collnearty-reducng observatons. References 1 D. C. Montgomery, E. A. Peck, and G. G. Vvng, Introducton to Lnear Regresson Analyss, John Wley & Sons, New York, NY, USA, 3rd edton, Md. Kamruzzaman and A. H. M. R. Imon, Hgh leverage pont: another source of multcollnearty, Pakstan Journal of Statstcs, vol. 18, no. 3, pp , M. H. Kutner, C. J. Nachtshem, and J. Neter, Appled Lnear Regresson Models, McGraw-Hll, New York, NY, USA, S. Chatterjee and A. S. Had, Regresson Analyss by Examples, John Wley & Sons, New York, NY, USA, 4th edton, A. S. Had, Dagnosng collnearty-nfluental observatons, Computatonal Statstcs & Data Analyss, vol. 7, no. 2, pp , M, A. Habshah, Bagher, A. H. M. R, and Imon, The applcaton of robust multcollnearty dagnostc method based on robust coeffcent determnaton to a non-collnear data, Journal of Appled Scences, vol. 1, no. 8, pp , M. Habshah, A. Bagher, and A. H. M. R. Imon, Hgh leverage collnearty-enhancng observatons and ts effect on multcollnearty pattern; Monte Carlo smulaton study, Sans Malaysana, vol. 4, no. 12, pp , A. Bagher, H. Md, and A. H. M. R. Imon, The effect of collnearty-nfluental observatons on collnear data set: A monte carlo smulaton study, Journal of Appled Scences, vol. 1, no. 18, pp , A. Bagher and H. Md, On the performance of robust varance nflaton factors, Internatonal Journal of Agrcultural and Statstcal Scences, vol. 7, no. 1, pp , A. Bagher, M. Habshah, and R. H. M. R. Imon, A novel collnearty-nfluental observaton dagnostc measure based on a group deleton approach, Communcatons n Statstcs, vol. 41, no. 8, pp , M. Habshah, M. R. Norazan, and A. H. M. R. Imon, The performance of dagnostc-robust generalzed potentals for the dentfcaton of multple hgh leverage ponts n lnear regresson, Journal of Appled Statstcs, vol. 36, no. 5-6, pp , D. Sengupta and P. Bhmasankaram, On the roles of observatons n collnearty n the lnear model, Journal of the Amercan Statstcal Assocaton, vol. 92, no. 439, pp , 1997.

16 16 Mathematcal Problems n Engneerng 13 A. S. Had, A new measure of overall potental nfluence n lnear regresson, Computatonal Statstcs and Data Analyss, vol. 14, no. 1, pp. 1 27, A. H. M. R. Imon, Identfyng multple hgh leverage ponts n lnear regresson, Journal of Statstcal Studes, vol. 3, pp , 22, Specal Volume n Honour of Professor Mr Masoom Al. 15 D. A. Belsley, E. Kuh, and R. E. Welsch, Regresson Dagnostcs: Identfyng Influental Data and Sources of Collnearty, John Wley & Sons, New York, NY, USA, D. A. Belsley, Condtonng Dagnostcs-Collnearty and Weak Data n Regresson, John Wley & Sons, New York, NY, USA, D. C. Hoagln and R. E. Welsch, The Hat Matrx n regresson and ANOVA, Journal of the Amercan Statstcal Assocaton, vol. 32, no. 1, pp , P. Rousseeuw, Multvarate estmaton wth hgh breakdown pont, n Mathematcal Statstcs and Applcatons, pp , Redel, Dordrecht, The Netherlands, D. C. Montgomery, G. C. Runger, and N. F. Hubele, Engneerng Statstcs, John Wley & Sons, New York, NY, USA, 5nd edton, G. W. Stewart, Collnearty and least squares regresson, Statstcal Scence, vol. 2, no. 1, pp. 68 1, K. D. Lawrence and J. L. Arthur, Robust Regresson; Analyss and Applcatons, Marcel Dekker, New York, NY, USA, 199.

17 Advances n Operatons Research Hndaw Publshng Corporaton Advances n Decson Scences Hndaw Publshng Corporaton Journal of Appled Mathematcs Algebra Hndaw Publshng Corporaton Hndaw Publshng Corporaton Journal of Probablty and Statstcs The Scentfc World Journal Hndaw Publshng Corporaton Hndaw Publshng Corporaton Internatonal Journal of Dfferental Equatons Hndaw Publshng Corporaton Submt your manuscrpts at Internatonal Journal of Advances n Combnatorcs Hndaw Publshng Corporaton Mathematcal Physcs Hndaw Publshng Corporaton Journal of Complex Analyss Hndaw Publshng Corporaton Internatonal Journal of Mathematcs and Mathematcal Scences Mathematcal Problems n Engneerng Journal of Mathematcs Hndaw Publshng Corporaton Hndaw Publshng Corporaton Hndaw Publshng Corporaton Dscrete Mathematcs Journal of Hndaw Publshng Corporaton Dscrete Dynamcs n Nature and Socety Journal of Functon Spaces Hndaw Publshng Corporaton Abstract and Appled Analyss Hndaw Publshng Corporaton Hndaw Publshng Corporaton Internatonal Journal of Journal of Stochastc Analyss Optmzaton Hndaw Publshng Corporaton Hndaw Publshng Corporaton

Robust Estimations as a Remedy for Multicollinearity Caused by Multiple High Leverage Points

Robust Estimations as a Remedy for Multicollinearity Caused by Multiple High Leverage Points Journal of Mathematcs and Statstcs 5 (4): 311-31, 009 ISSN 1549-3644 009 Scence Publcatons Robust Estmatons as a Remedy for Multcollnearty Caused by Multple Hgh Leverage Ponts Arezoo Bagher and Habshah

More information

Psychology 282 Lecture #24 Outline Regression Diagnostics: Outliers

Psychology 282 Lecture #24 Outline Regression Diagnostics: Outliers Psychology 282 Lecture #24 Outlne Regresson Dagnostcs: Outlers In an earler lecture we studed the statstcal assumptons underlyng the regresson model, ncludng the followng ponts: Formal statement of assumptons.

More information

Research Article Green s Theorem for Sign Data

Research Article Green s Theorem for Sign Data Internatonal Scholarly Research Network ISRN Appled Mathematcs Volume 2012, Artcle ID 539359, 10 pages do:10.5402/2012/539359 Research Artcle Green s Theorem for Sgn Data Lous M. Houston The Unversty of

More information

Comparison of the Population Variance Estimators. of 2-Parameter Exponential Distribution Based on. Multiple Criteria Decision Making Method

Comparison of the Population Variance Estimators. of 2-Parameter Exponential Distribution Based on. Multiple Criteria Decision Making Method Appled Mathematcal Scences, Vol. 7, 0, no. 47, 07-0 HIARI Ltd, www.m-hkar.com Comparson of the Populaton Varance Estmators of -Parameter Exponental Dstrbuton Based on Multple Crtera Decson Makng Method

More information

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 30 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 2 Remedes for multcollnearty Varous technques have

More information

Robust Logistic Ridge Regression Estimator in the Presence of High Leverage Multicollinear Observations

Robust Logistic Ridge Regression Estimator in the Presence of High Leverage Multicollinear Observations Mathematcal and Computatonal Methods n Scence and Engneerng Robust Logstc Rdge Regresson Estmator n the Presence of Hgh Leverage Multcollnear Observatons SYAIBA BALQISH ARIFFIN 1 AND HABSHAH MIDI 1, Faculty

More information

Estimation Methods for Multicollinearity Proplem Combined with High Leverage Data Points

Estimation Methods for Multicollinearity Proplem Combined with High Leverage Data Points Journal of Mathematcs and Statstcs 7 (): 19-136, 011 ISSN 1549-3644 010 Scence Publcatons Estmaton Methods for Multcollnearty Proplem Combned wth Hgh Leverage Data Ponts Moawad El-Fallah and Abd El-Sallam

More information

Comparison of Regression Lines

Comparison of Regression Lines STATGRAPHICS Rev. 9/13/2013 Comparson of Regresson Lnes Summary... 1 Data Input... 3 Analyss Summary... 4 Plot of Ftted Model... 6 Condtonal Sums of Squares... 6 Analyss Optons... 7 Forecasts... 8 Confdence

More information

A Robust Method for Calculating the Correlation Coefficient

A Robust Method for Calculating the Correlation Coefficient A Robust Method for Calculatng the Correlaton Coeffcent E.B. Nven and C. V. Deutsch Relatonshps between prmary and secondary data are frequently quantfed usng the correlaton coeffcent; however, the tradtonal

More information

Journal of Modern Applied Statistical Methods

Journal of Modern Applied Statistical Methods Journal of Modern Appled Statstcal Methods Volume 17 Issue 1 Artcle 17 6-29-2018 Robust Heteroscedastcty Consstent Covarance Matrx Estmator based on Robust Mahalanobs Dstance and Dagnostc Robust Generalzed

More information

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4)

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4) I. Classcal Assumptons Econ7 Appled Econometrcs Topc 3: Classcal Model (Studenmund, Chapter 4) We have defned OLS and studed some algebrac propertes of OLS. In ths topc we wll study statstcal propertes

More information

The Order Relation and Trace Inequalities for. Hermitian Operators

The Order Relation and Trace Inequalities for. Hermitian Operators Internatonal Mathematcal Forum, Vol 3, 08, no, 507-57 HIKARI Ltd, wwwm-hkarcom https://doorg/0988/mf088055 The Order Relaton and Trace Inequaltes for Hermtan Operators Y Huang School of Informaton Scence

More information

[ ] λ λ λ. Multicollinearity. multicollinearity Ragnar Frisch (1934) perfect exact. collinearity. multicollinearity. exact

[ ] λ λ λ. Multicollinearity. multicollinearity Ragnar Frisch (1934) perfect exact. collinearity. multicollinearity. exact Multcollnearty multcollnearty Ragnar Frsch (934 perfect exact collnearty multcollnearty K exact λ λ λ K K x+ x+ + x 0 0.. λ, λ, λk 0 0.. x perfect ntercorrelated λ λ λ x+ x+ + KxK + v 0 0.. v 3 y β + β

More information

Chapter 15 Student Lecture Notes 15-1

Chapter 15 Student Lecture Notes 15-1 Chapter 15 Student Lecture Notes 15-1 Basc Busness Statstcs (9 th Edton) Chapter 15 Multple Regresson Model Buldng 004 Prentce-Hall, Inc. Chap 15-1 Chapter Topcs The Quadratc Regresson Model Usng Transformatons

More information

Statistical inference for generalized Pareto distribution based on progressive Type-II censored data with random removals

Statistical inference for generalized Pareto distribution based on progressive Type-II censored data with random removals Internatonal Journal of Scentfc World, 2 1) 2014) 1-9 c Scence Publshng Corporaton www.scencepubco.com/ndex.php/ijsw do: 10.14419/jsw.v21.1780 Research Paper Statstcal nference for generalzed Pareto dstrbuton

More information

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur Analyss of Varance and Desgn of Experment-I MODULE VII LECTURE - 3 ANALYSIS OF COVARIANCE Dr Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur Any scentfc experment s performed

More information

Negative Binomial Regression

Negative Binomial Regression STATGRAPHICS Rev. 9/16/2013 Negatve Bnomal Regresson Summary... 1 Data Input... 3 Statstcal Model... 3 Analyss Summary... 4 Analyss Optons... 7 Plot of Ftted Model... 8 Observed Versus Predcted... 10 Predctons...

More information

Chapter 13: Multiple Regression

Chapter 13: Multiple Regression Chapter 13: Multple Regresson 13.1 Developng the multple-regresson Model The general model can be descrbed as: It smplfes for two ndependent varables: The sample ft parameter b 0, b 1, and b are used to

More information

Influence Diagnostics on Competing Risks Using Cox s Model with Censored Data. Jalan Gombak, 53100, Kuala Lumpur, Malaysia.

Influence Diagnostics on Competing Risks Using Cox s Model with Censored Data. Jalan Gombak, 53100, Kuala Lumpur, Malaysia. Proceedngs of the 8th WSEAS Internatonal Conference on APPLIED MAHEMAICS, enerfe, Span, December 16-18, 5 (pp14-138) Influence Dagnostcs on Competng Rsks Usng Cox s Model wth Censored Data F. A. M. Elfak

More information

On the Influential Points in the Functional Circular Relationship Models

On the Influential Points in the Functional Circular Relationship Models On the Influental Ponts n the Functonal Crcular Relatonshp Models Department of Mathematcs, Faculty of Scence Al-Azhar Unversty-Gaza, Gaza, Palestne alzad33@yahoo.com Abstract If the nterest s to calbrate

More information

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS Avalable onlne at http://sck.org J. Math. Comput. Sc. 3 (3), No., 6-3 ISSN: 97-537 COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

More information

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction ECONOMICS 5* -- NOTE (Summary) ECON 5* -- NOTE The Multple Classcal Lnear Regresson Model (CLRM): Specfcaton and Assumptons. Introducton CLRM stands for the Classcal Lnear Regresson Model. The CLRM s also

More information

on the improved Partial Least Squares regression

on the improved Partial Least Squares regression Internatonal Conference on Manufacturng Scence and Engneerng (ICMSE 05) Identfcaton of the multvarable outlers usng T eclpse chart based on the mproved Partal Least Squares regresson Lu Yunlan,a X Yanhu,b

More information

Chapter 12 Analysis of Covariance

Chapter 12 Analysis of Covariance Chapter Analyss of Covarance Any scentfc experment s performed to know somethng that s unknown about a group of treatments and to test certan hypothess about the correspondng treatment effect When varablty

More information

On the detection of influential outliers in linear regression analysis

On the detection of influential outliers in linear regression analysis Amercan Journal of Theoretcal and Appled Statstcs 04; 3(4): 00-06 Publshed onlne July 30, 04 (http://www.scencepublshnggroup.com/j/ajtas) do: 0.648/j.ajtas.040304.4 ISSN: 36-8999 (Prnt); ISSN: 36-9006

More information

Global Sensitivity. Tuesday 20 th February, 2018

Global Sensitivity. Tuesday 20 th February, 2018 Global Senstvty Tuesday 2 th February, 28 ) Local Senstvty Most senstvty analyses [] are based on local estmates of senstvty, typcally by expandng the response n a Taylor seres about some specfc values

More information

Number of cases Number of factors Number of covariates Number of levels of factor i. Value of the dependent variable for case k

Number of cases Number of factors Number of covariates Number of levels of factor i. Value of the dependent variable for case k ANOVA Model and Matrx Computatons Notaton The followng notaton s used throughout ths chapter unless otherwse stated: N F CN Y Z j w W Number of cases Number of factors Number of covarates Number of levels

More information

Chapter 9: Statistical Inference and the Relationship between Two Variables

Chapter 9: Statistical Inference and the Relationship between Two Variables Chapter 9: Statstcal Inference and the Relatonshp between Two Varables Key Words The Regresson Model The Sample Regresson Equaton The Pearson Correlaton Coeffcent Learnng Outcomes After studyng ths chapter,

More information

x i1 =1 for all i (the constant ).

x i1 =1 for all i (the constant ). Chapter 5 The Multple Regresson Model Consder an economc model where the dependent varable s a functon of K explanatory varables. The economc model has the form: y = f ( x,x,..., ) xk Approxmate ths by

More information

Ridge Regression Estimators with the Problem. of Multicollinearity

Ridge Regression Estimators with the Problem. of Multicollinearity Appled Mathematcal Scences, Vol. 7, 2013, no. 50, 2469-2480 HIKARI Ltd, www.m-hkar.com Rdge Regresson Estmators wth the Problem of Multcollnearty Mae M. Kamel Statstc Department, Faculty of Commerce Tanta

More information

Econ107 Applied Econometrics Topic 9: Heteroskedasticity (Studenmund, Chapter 10)

Econ107 Applied Econometrics Topic 9: Heteroskedasticity (Studenmund, Chapter 10) I. Defnton and Problems Econ7 Appled Econometrcs Topc 9: Heteroskedastcty (Studenmund, Chapter ) We now relax another classcal assumpton. Ths s a problem that arses often wth cross sectons of ndvduals,

More information

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur Analyss of Varance and Desgn of Exerments-I MODULE III LECTURE - 2 EXPERIMENTAL DESIGN MODELS Dr. Shalabh Deartment of Mathematcs and Statstcs Indan Insttute of Technology Kanur 2 We consder the models

More information

2016 Wiley. Study Session 2: Ethical and Professional Standards Application

2016 Wiley. Study Session 2: Ethical and Professional Standards Application 6 Wley Study Sesson : Ethcal and Professonal Standards Applcaton LESSON : CORRECTION ANALYSIS Readng 9: Correlaton and Regresson LOS 9a: Calculate and nterpret a sample covarance and a sample correlaton

More information

Statistics for Economics & Business

Statistics for Economics & Business Statstcs for Economcs & Busness Smple Lnear Regresson Learnng Objectves In ths chapter, you learn: How to use regresson analyss to predct the value of a dependent varable based on an ndependent varable

More information

LINEAR REGRESSION ANALYSIS. MODULE VIII Lecture Indicator Variables

LINEAR REGRESSION ANALYSIS. MODULE VIII Lecture Indicator Variables LINEAR REGRESSION ANALYSIS MODULE VIII Lecture - 7 Indcator Varables Dr. Shalabh Department of Maematcs and Statstcs Indan Insttute of Technology Kanpur Indcator varables versus quanttatve explanatory

More information

Research Article Global Sufficient Optimality Conditions for a Special Cubic Minimization Problem

Research Article Global Sufficient Optimality Conditions for a Special Cubic Minimization Problem Mathematcal Problems n Engneerng Volume 2012, Artcle ID 871741, 16 pages do:10.1155/2012/871741 Research Artcle Global Suffcent Optmalty Condtons for a Specal Cubc Mnmzaton Problem Xaome Zhang, 1 Yanjun

More information

Case Study of Markov Chains Ray-Knight Compactification

Case Study of Markov Chains Ray-Knight Compactification Internatonal Journal of Contemporary Mathematcal Scences Vol. 9, 24, no. 6, 753-76 HIKAI Ltd, www.m-har.com http://dx.do.org/.2988/cms.24.46 Case Study of Marov Chans ay-knght Compactfcaton HaXa Du and

More information

Department of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6

Department of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6 Department of Quanttatve Methods & Informaton Systems Tme Seres and Ther Components QMIS 30 Chapter 6 Fall 00 Dr. Mohammad Zanal These sldes were modfed from ther orgnal source for educatonal purpose only.

More information

The Study of Teaching-learning-based Optimization Algorithm

The Study of Teaching-learning-based Optimization Algorithm Advanced Scence and Technology Letters Vol. (AST 06), pp.05- http://dx.do.org/0.57/astl.06. The Study of Teachng-learnng-based Optmzaton Algorthm u Sun, Yan fu, Lele Kong, Haolang Q,, Helongang Insttute

More information

Chapter 8 Indicator Variables

Chapter 8 Indicator Variables Chapter 8 Indcator Varables In general, e explanatory varables n any regresson analyss are assumed to be quanttatve n nature. For example, e varables lke temperature, dstance, age etc. are quanttatve n

More information

On the correction of the h-index for career length

On the correction of the h-index for career length 1 On the correcton of the h-ndex for career length by L. Egghe Unverstet Hasselt (UHasselt), Campus Depenbeek, Agoralaan, B-3590 Depenbeek, Belgum 1 and Unverstet Antwerpen (UA), IBW, Stadscampus, Venusstraat

More information

Turbulence classification of load data by the frequency and severity of wind gusts. Oscar Moñux, DEWI GmbH Kevin Bleibler, DEWI GmbH

Turbulence classification of load data by the frequency and severity of wind gusts. Oscar Moñux, DEWI GmbH Kevin Bleibler, DEWI GmbH Turbulence classfcaton of load data by the frequency and severty of wnd gusts Introducton Oscar Moñux, DEWI GmbH Kevn Blebler, DEWI GmbH Durng the wnd turbne developng process, one of the most mportant

More information

Liu-type Negative Binomial Regression: A Comparison of Recent Estimators and Applications

Liu-type Negative Binomial Regression: A Comparison of Recent Estimators and Applications Lu-type Negatve Bnomal Regresson: A Comparson of Recent Estmators and Applcatons Yasn Asar Department of Mathematcs-Computer Scences, Necmettn Erbaan Unversty, Konya 4090, Turey, yasar@onya.edu.tr, yasnasar@hotmal.com

More information

Lecture 6: Introduction to Linear Regression

Lecture 6: Introduction to Linear Regression Lecture 6: Introducton to Lnear Regresson An Manchakul amancha@jhsph.edu 24 Aprl 27 Lnear regresson: man dea Lnear regresson can be used to study an outcome as a lnear functon of a predctor Example: 6

More information

Linear Regression Analysis: Terminology and Notation

Linear Regression Analysis: Terminology and Notation ECON 35* -- Secton : Basc Concepts of Regresson Analyss (Page ) Lnear Regresson Analyss: Termnology and Notaton Consder the generc verson of the smple (two-varable) lnear regresson model. It s represented

More information

REGRESSION ANALYSIS II- MULTICOLLINEARITY

REGRESSION ANALYSIS II- MULTICOLLINEARITY REGRESSION ANALYSIS II- MULTICOLLINEARITY QUESTION 1 Departments of Open Unversty of Cyprus A and B consst of na = 35 and nb = 30 students respectvely. The students of department A acheved an average test

More information

LECTURE 9 CANONICAL CORRELATION ANALYSIS

LECTURE 9 CANONICAL CORRELATION ANALYSIS LECURE 9 CANONICAL CORRELAION ANALYSIS Introducton he concept of canoncal correlaton arses when we want to quantfy the assocatons between two sets of varables. For example, suppose that the frst set of

More information

Convexity preserving interpolation by splines of arbitrary degree

Convexity preserving interpolation by splines of arbitrary degree Computer Scence Journal of Moldova, vol.18, no.1(52), 2010 Convexty preservng nterpolaton by splnes of arbtrary degree Igor Verlan Abstract In the present paper an algorthm of C 2 nterpolaton of dscrete

More information

Sharp integral inequalities involving high-order partial derivatives. Journal Of Inequalities And Applications, 2008, v. 2008, article no.

Sharp integral inequalities involving high-order partial derivatives. Journal Of Inequalities And Applications, 2008, v. 2008, article no. Ttle Sharp ntegral nequaltes nvolvng hgh-order partal dervatves Authors Zhao, CJ; Cheung, WS Ctaton Journal Of Inequaltes And Applcatons, 008, v. 008, artcle no. 5747 Issued Date 008 URL http://hdl.handle.net/07/569

More information

Simulated Power of the Discrete Cramér-von Mises Goodness-of-Fit Tests

Simulated Power of the Discrete Cramér-von Mises Goodness-of-Fit Tests Smulated of the Cramér-von Mses Goodness-of-Ft Tests Steele, M., Chaselng, J. and 3 Hurst, C. School of Mathematcal and Physcal Scences, James Cook Unversty, Australan School of Envronmental Studes, Grffth

More information

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems Numercal Analyss by Dr. Anta Pal Assstant Professor Department of Mathematcs Natonal Insttute of Technology Durgapur Durgapur-713209 emal: anta.bue@gmal.com 1 . Chapter 5 Soluton of System of Lnear Equatons

More information

Primer on High-Order Moment Estimators

Primer on High-Order Moment Estimators Prmer on Hgh-Order Moment Estmators Ton M. Whted July 2007 The Errors-n-Varables Model We wll start wth the classcal EIV for one msmeasured regressor. The general case s n Erckson and Whted Econometrc

More information

Chapter 15 - Multiple Regression

Chapter 15 - Multiple Regression Chapter - Multple Regresson Chapter - Multple Regresson Multple Regresson Model The equaton that descrbes how the dependent varable y s related to the ndependent varables x, x,... x p and an error term

More information

Durban Watson for Testing the Lack-of-Fit of Polynomial Regression Models without Replications

Durban Watson for Testing the Lack-of-Fit of Polynomial Regression Models without Replications Durban Watson for Testng the Lack-of-Ft of Polynomal Regresson Models wthout Replcatons Ruba A. Alyaf, Maha A. Omar, Abdullah A. Al-Shha ralyaf@ksu.edu.sa, maomar@ksu.edu.sa, aalshha@ksu.edu.sa Department

More information

The Minimum Universal Cost Flow in an Infeasible Flow Network

The Minimum Universal Cost Flow in an Infeasible Flow Network Journal of Scences, Islamc Republc of Iran 17(2): 175-180 (2006) Unversty of Tehran, ISSN 1016-1104 http://jscencesutacr The Mnmum Unversal Cost Flow n an Infeasble Flow Network H Saleh Fathabad * M Bagheran

More information

Kernel Methods and SVMs Extension

Kernel Methods and SVMs Extension Kernel Methods and SVMs Extenson The purpose of ths document s to revew materal covered n Machne Learnng 1 Supervsed Learnng regardng support vector machnes (SVMs). Ths document also provdes a general

More information

2E Pattern Recognition Solutions to Introduction to Pattern Recognition, Chapter 2: Bayesian pattern classification

2E Pattern Recognition Solutions to Introduction to Pattern Recognition, Chapter 2: Bayesian pattern classification E395 - Pattern Recognton Solutons to Introducton to Pattern Recognton, Chapter : Bayesan pattern classfcaton Preface Ths document s a soluton manual for selected exercses from Introducton to Pattern Recognton

More information

Statistics II Final Exam 26/6/18

Statistics II Final Exam 26/6/18 Statstcs II Fnal Exam 26/6/18 Academc Year 2017/18 Solutons Exam duraton: 2 h 30 mn 1. (3 ponts) A town hall s conductng a study to determne the amount of leftover food produced by the restaurants n the

More information

ANOMALIES OF THE MAGNITUDE OF THE BIAS OF THE MAXIMUM LIKELIHOOD ESTIMATOR OF THE REGRESSION SLOPE

ANOMALIES OF THE MAGNITUDE OF THE BIAS OF THE MAXIMUM LIKELIHOOD ESTIMATOR OF THE REGRESSION SLOPE P a g e ANOMALIES OF THE MAGNITUDE OF THE BIAS OF THE MAXIMUM LIKELIHOOD ESTIMATOR OF THE REGRESSION SLOPE Darmud O Drscoll ¹, Donald E. Ramrez ² ¹ Head of Department of Mathematcs and Computer Studes

More information

Asymptotics of the Solution of a Boundary Value. Problem for One-Characteristic Differential. Equation Degenerating into a Parabolic Equation

Asymptotics of the Solution of a Boundary Value. Problem for One-Characteristic Differential. Equation Degenerating into a Parabolic Equation Nonl. Analyss and Dfferental Equatons, ol., 4, no., 5 - HIKARI Ltd, www.m-har.com http://dx.do.org/.988/nade.4.456 Asymptotcs of the Soluton of a Boundary alue Problem for One-Characterstc Dfferental Equaton

More information

Statistics for Business and Economics

Statistics for Business and Economics Statstcs for Busness and Economcs Chapter 11 Smple Regresson Copyrght 010 Pearson Educaton, Inc. Publshng as Prentce Hall Ch. 11-1 11.1 Overvew of Lnear Models n An equaton can be ft to show the best lnear

More information

arxiv:cs.cv/ Jun 2000

arxiv:cs.cv/ Jun 2000 Correlaton over Decomposed Sgnals: A Non-Lnear Approach to Fast and Effectve Sequences Comparson Lucano da Fontoura Costa arxv:cs.cv/0006040 28 Jun 2000 Cybernetc Vson Research Group IFSC Unversty of São

More information

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 31 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 6. Rdge regresson The OLSE s the best lnear unbased

More information

Existence of Two Conjugate Classes of A 5 within S 6. by Use of Character Table of S 6

Existence of Two Conjugate Classes of A 5 within S 6. by Use of Character Table of S 6 Internatonal Mathematcal Forum, Vol. 8, 2013, no. 32, 1591-159 HIKARI Ltd, www.m-hkar.com http://dx.do.org/10.12988/mf.2013.3359 Exstence of Two Conjugate Classes of A 5 wthn S by Use of Character Table

More information

Formulas for the Determinant

Formulas for the Determinant page 224 224 CHAPTER 3 Determnants e t te t e 2t 38 A = e t 2te t e 2t e t te t 2e 2t 39 If 123 A = 345, 456 compute the matrx product A adj(a) What can you conclude about det(a)? For Problems 40 43, use

More information

BOOTSTRAP METHOD FOR TESTING OF EQUALITY OF SEVERAL MEANS. M. Krishna Reddy, B. Naveen Kumar and Y. Ramu

BOOTSTRAP METHOD FOR TESTING OF EQUALITY OF SEVERAL MEANS. M. Krishna Reddy, B. Naveen Kumar and Y. Ramu BOOTSTRAP METHOD FOR TESTING OF EQUALITY OF SEVERAL MEANS M. Krshna Reddy, B. Naveen Kumar and Y. Ramu Department of Statstcs, Osmana Unversty, Hyderabad -500 007, Inda. nanbyrozu@gmal.com, ramu0@gmal.com

More information

A PROBABILITY-DRIVEN SEARCH ALGORITHM FOR SOLVING MULTI-OBJECTIVE OPTIMIZATION PROBLEMS

A PROBABILITY-DRIVEN SEARCH ALGORITHM FOR SOLVING MULTI-OBJECTIVE OPTIMIZATION PROBLEMS HCMC Unversty of Pedagogy Thong Nguyen Huu et al. A PROBABILITY-DRIVEN SEARCH ALGORITHM FOR SOLVING MULTI-OBJECTIVE OPTIMIZATION PROBLEMS Thong Nguyen Huu and Hao Tran Van Department of mathematcs-nformaton,

More information

A new construction of 3-separable matrices via an improved decoding of Macula s construction

A new construction of 3-separable matrices via an improved decoding of Macula s construction Dscrete Optmzaton 5 008 700 704 Contents lsts avalable at ScenceDrect Dscrete Optmzaton journal homepage: wwwelsevercom/locate/dsopt A new constructon of 3-separable matrces va an mproved decodng of Macula

More information

Markov Chain Monte Carlo Lecture 6

Markov Chain Monte Carlo Lecture 6 where (x 1,..., x N ) X N, N s called the populaton sze, f(x) f (x) for at least one {1, 2,..., N}, and those dfferent from f(x) are called the tral dstrbutons n terms of mportance samplng. Dfferent ways

More information

DERIVATION OF THE PROBABILITY PLOT CORRELATION COEFFICIENT TEST STATISTICS FOR THE GENERALIZED LOGISTIC DISTRIBUTION

DERIVATION OF THE PROBABILITY PLOT CORRELATION COEFFICIENT TEST STATISTICS FOR THE GENERALIZED LOGISTIC DISTRIBUTION Internatonal Worshop ADVANCES IN STATISTICAL HYDROLOGY May 3-5, Taormna, Italy DERIVATION OF THE PROBABILITY PLOT CORRELATION COEFFICIENT TEST STATISTICS FOR THE GENERALIZED LOGISTIC DISTRIBUTION by Sooyoung

More information

ISSN: ISO 9001:2008 Certified International Journal of Engineering and Innovative Technology (IJEIT) Volume 3, Issue 1, July 2013

ISSN: ISO 9001:2008 Certified International Journal of Engineering and Innovative Technology (IJEIT) Volume 3, Issue 1, July 2013 ISSN: 2277-375 Constructon of Trend Free Run Orders for Orthogonal rrays Usng Codes bstract: Sometmes when the expermental runs are carred out n a tme order sequence, the response can depend on the run

More information

RESIDUALS AND INFLUENCE IN NONLINEAR REGRESSION FOR REPEATED MEASUREMENT DATA

RESIDUALS AND INFLUENCE IN NONLINEAR REGRESSION FOR REPEATED MEASUREMENT DATA Operatons Research and Applcatons : An Internatonal Journal (ORAJ), Vol.4, No.3/4, November 17 RESIDUALS AND INFLUENCE IN NONLINEAR REGRESSION FOR REPEAED MEASUREMEN DAA Munsr Al, Yu Feng, Al choo, Zamr

More information

Using T.O.M to Estimate Parameter of distributions that have not Single Exponential Family

Using T.O.M to Estimate Parameter of distributions that have not Single Exponential Family IOSR Journal of Mathematcs IOSR-JM) ISSN: 2278-5728. Volume 3, Issue 3 Sep-Oct. 202), PP 44-48 www.osrjournals.org Usng T.O.M to Estmate Parameter of dstrbutons that have not Sngle Exponental Famly Jubran

More information

Basically, if you have a dummy dependent variable you will be estimating a probability.

Basically, if you have a dummy dependent variable you will be estimating a probability. ECON 497: Lecture Notes 13 Page 1 of 1 Metropoltan State Unversty ECON 497: Research and Forecastng Lecture Notes 13 Dummy Dependent Varable Technques Studenmund Chapter 13 Bascally, f you have a dummy

More information

FREQUENCY DISTRIBUTIONS Page 1 of The idea of a frequency distribution for sets of observations will be introduced,

FREQUENCY DISTRIBUTIONS Page 1 of The idea of a frequency distribution for sets of observations will be introduced, FREQUENCY DISTRIBUTIONS Page 1 of 6 I. Introducton 1. The dea of a frequency dstrbuton for sets of observatons wll be ntroduced, together wth some of the mechancs for constructng dstrbutons of data. Then

More information

Chap 10: Diagnostics, p384

Chap 10: Diagnostics, p384 Chap 10: Dagnostcs, p384 Multcollnearty 10.5 p406 Defnton Multcollnearty exsts when two or more ndependent varables used n regresson are moderately or hghly correlated. - when multcollnearty exsts, regresson

More information

Uncertainty and auto-correlation in. Measurement

Uncertainty and auto-correlation in. Measurement Uncertanty and auto-correlaton n arxv:1707.03276v2 [physcs.data-an] 30 Dec 2017 Measurement Markus Schebl Federal Offce of Metrology and Surveyng (BEV), 1160 Venna, Austra E-mal: markus.schebl@bev.gv.at

More information

Notes on Frequency Estimation in Data Streams

Notes on Frequency Estimation in Data Streams Notes on Frequency Estmaton n Data Streams In (one of) the data streamng model(s), the data s a sequence of arrvals a 1, a 2,..., a m of the form a j = (, v) where s the dentty of the tem and belongs to

More information

Testing for outliers in nonlinear longitudinal data models based on M-estimation

Testing for outliers in nonlinear longitudinal data models based on M-estimation ISS 1746-7659, England, UK Journal of Informaton and Computng Scence Vol 1, o, 017, pp107-11 estng for outlers n nonlnear longtudnal data models based on M-estmaton Huhu Sun 1 1 School of Mathematcs and

More information

Exponential Type Product Estimator for Finite Population Mean with Information on Auxiliary Attribute

Exponential Type Product Estimator for Finite Population Mean with Information on Auxiliary Attribute Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 193-9466 Vol. 10, Issue 1 (June 015), pp. 106-113 Applcatons and Appled Mathematcs: An Internatonal Journal (AAM) Exponental Tpe Product Estmator

More information

See Book Chapter 11 2 nd Edition (Chapter 10 1 st Edition)

See Book Chapter 11 2 nd Edition (Chapter 10 1 st Edition) Count Data Models See Book Chapter 11 2 nd Edton (Chapter 10 1 st Edton) Count data consst of non-negatve nteger values Examples: number of drver route changes per week, the number of trp departure changes

More information

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number

More information

A Hybrid Variational Iteration Method for Blasius Equation

A Hybrid Variational Iteration Method for Blasius Equation Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 1932-9466 Vol. 10, Issue 1 (June 2015), pp. 223-229 Applcatons and Appled Mathematcs: An Internatonal Journal (AAM) A Hybrd Varatonal Iteraton Method

More information

UNIVERSITY OF TORONTO Faculty of Arts and Science. December 2005 Examinations STA437H1F/STA1005HF. Duration - 3 hours

UNIVERSITY OF TORONTO Faculty of Arts and Science. December 2005 Examinations STA437H1F/STA1005HF. Duration - 3 hours UNIVERSITY OF TORONTO Faculty of Arts and Scence December 005 Examnatons STA47HF/STA005HF Duraton - hours AIDS ALLOWED: (to be suppled by the student) Non-programmable calculator One handwrtten 8.5'' x

More information

Chapter 11: Simple Linear Regression and Correlation

Chapter 11: Simple Linear Regression and Correlation Chapter 11: Smple Lnear Regresson and Correlaton 11-1 Emprcal Models 11-2 Smple Lnear Regresson 11-3 Propertes of the Least Squares Estmators 11-4 Hypothess Test n Smple Lnear Regresson 11-4.1 Use of t-tests

More information

DETERMINATION OF TEMPERATURE DISTRIBUTION FOR ANNULAR FINS WITH TEMPERATURE DEPENDENT THERMAL CONDUCTIVITY BY HPM

DETERMINATION OF TEMPERATURE DISTRIBUTION FOR ANNULAR FINS WITH TEMPERATURE DEPENDENT THERMAL CONDUCTIVITY BY HPM Ganj, Z. Z., et al.: Determnaton of Temperature Dstrbuton for S111 DETERMINATION OF TEMPERATURE DISTRIBUTION FOR ANNULAR FINS WITH TEMPERATURE DEPENDENT THERMAL CONDUCTIVITY BY HPM by Davood Domr GANJI

More information

DO NOT OPEN THE QUESTION PAPER UNTIL INSTRUCTED TO DO SO BY THE CHIEF INVIGILATOR. Introductory Econometrics 1 hour 30 minutes

DO NOT OPEN THE QUESTION PAPER UNTIL INSTRUCTED TO DO SO BY THE CHIEF INVIGILATOR. Introductory Econometrics 1 hour 30 minutes 25/6 Canddates Only January Examnatons 26 Student Number: Desk Number:...... DO NOT OPEN THE QUESTION PAPER UNTIL INSTRUCTED TO DO SO BY THE CHIEF INVIGILATOR Department Module Code Module Ttle Exam Duraton

More information

On the Interval Zoro Symmetric Single-step Procedure for Simultaneous Finding of Polynomial Zeros

On the Interval Zoro Symmetric Single-step Procedure for Simultaneous Finding of Polynomial Zeros Appled Mathematcal Scences, Vol. 5, 2011, no. 75, 3693-3706 On the Interval Zoro Symmetrc Sngle-step Procedure for Smultaneous Fndng of Polynomal Zeros S. F. M. Rusl, M. Mons, M. A. Hassan and W. J. Leong

More information

ECONOMICS 351*-A Mid-Term Exam -- Fall Term 2000 Page 1 of 13 pages. QUEEN'S UNIVERSITY AT KINGSTON Department of Economics

ECONOMICS 351*-A Mid-Term Exam -- Fall Term 2000 Page 1 of 13 pages. QUEEN'S UNIVERSITY AT KINGSTON Department of Economics ECOOMICS 35*-A Md-Term Exam -- Fall Term 000 Page of 3 pages QUEE'S UIVERSITY AT KIGSTO Department of Economcs ECOOMICS 35* - Secton A Introductory Econometrcs Fall Term 000 MID-TERM EAM ASWERS MG Abbott

More information

January Examinations 2015

January Examinations 2015 24/5 Canddates Only January Examnatons 25 DO NOT OPEN THE QUESTION PAPER UNTIL INSTRUCTED TO DO SO BY THE CHIEF INVIGILATOR STUDENT CANDIDATE NO.. Department Module Code Module Ttle Exam Duraton (n words)

More information

Multivariate Ratio Estimator of the Population Total under Stratified Random Sampling

Multivariate Ratio Estimator of the Population Total under Stratified Random Sampling Open Journal of Statstcs, 0,, 300-304 ttp://dx.do.org/0.436/ojs.0.3036 Publsed Onlne July 0 (ttp://www.scrp.org/journal/ojs) Multvarate Rato Estmator of te Populaton Total under Stratfed Random Samplng

More information

International Journal of Engineering Research and Modern Education (IJERME) Impact Factor: 7.018, ISSN (Online): (

International Journal of Engineering Research and Modern Education (IJERME) Impact Factor: 7.018, ISSN (Online): ( CONSTRUCTION AND SELECTION OF CHAIN SAMPLING PLAN WITH ZERO INFLATED POISSON DISTRIBUTION A. Palansamy* & M. Latha** * Research Scholar, Department of Statstcs, Government Arts College, Udumalpet, Tamlnadu

More information

Non-Mixture Cure Model for Interval Censored Data: Simulation Study ABSTRACT

Non-Mixture Cure Model for Interval Censored Data: Simulation Study ABSTRACT Malaysan Journal of Mathematcal Scences 8(S): 37-44 (2014) Specal Issue: Internatonal Conference on Mathematcal Scences and Statstcs 2013 (ICMSS2013) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES Journal

More information

Methods of Detecting Outliers in A Regression Analysis Model.

Methods of Detecting Outliers in A Regression Analysis Model. Methods of Detectng Outlers n A Regresson Analyss Model. Ogu, A. I. *, Inyama, S. C+, Achugamonu, P. C++ *Department of Statstcs, Imo State Unversty,Owerr +Department of Mathematcs, Federal Unversty of

More information

Factor models with many assets: strong factors, weak factors, and the two-pass procedure

Factor models with many assets: strong factors, weak factors, and the two-pass procedure Factor models wth many assets: strong factors, weak factors, and the two-pass procedure Stanslav Anatolyev 1 Anna Mkusheva 2 1 CERGE-EI and NES 2 MIT December 2017 Stanslav Anatolyev and Anna Mkusheva

More information

Stochastic Restricted Maximum Likelihood Estimator in Logistic Regression Model

Stochastic Restricted Maximum Likelihood Estimator in Logistic Regression Model Open Journal of Statstcs, 05, 5, 837-85 Publshed Onlne December 05 n ScRes. http://www.scrp.org/journal/ojs http://dx.do.org/0.436/ojs.05.5708 Stochastc Restrcted Maxmum Lkelhood Estmator n Logstc Regresson

More information

Statistical Inference. 2.3 Summary Statistics Measures of Center and Spread. parameters ( population characteristics )

Statistical Inference. 2.3 Summary Statistics Measures of Center and Spread. parameters ( population characteristics ) Ismor Fscher, 8//008 Stat 54 / -8.3 Summary Statstcs Measures of Center and Spread Dstrbuton of dscrete contnuous POPULATION Random Varable, numercal True center =??? True spread =???? parameters ( populaton

More information

Power law and dimension of the maximum value for belief distribution with the max Deng entropy

Power law and dimension of the maximum value for belief distribution with the max Deng entropy Power law and dmenson of the maxmum value for belef dstrbuton wth the max Deng entropy Bngy Kang a, a College of Informaton Engneerng, Northwest A&F Unversty, Yanglng, Shaanx, 712100, Chna. Abstract Deng

More information

On an Extension of Stochastic Approximation EM Algorithm for Incomplete Data Problems. Vahid Tadayon 1

On an Extension of Stochastic Approximation EM Algorithm for Incomplete Data Problems. Vahid Tadayon 1 On an Extenson of Stochastc Approxmaton EM Algorthm for Incomplete Data Problems Vahd Tadayon Abstract: The Stochastc Approxmaton EM (SAEM algorthm, a varant stochastc approxmaton of EM, s a versatle tool

More information