The Single Pass Clustering Method. S. Rieber and V. P. Marathe. In information retrieval, several complex clustering methods exist

Size: px
Start display at page:

Download "The Single Pass Clustering Method. S. Rieber and V. P. Marathe. In information retrieval, several complex clustering methods exist"

Transcription

1 XIV- The Single ass Clustering Methd S. Rieber and V.. Marathe Abstract In infrmatin retrieval, several cmplex clustering methds exist which require extensive prcessing time and cmputer memry. A cheap clus- tering methd is develped which requires nly ne pass ver the dcument cllectin t generate clusters. The ne-pass clustering methd is investigated using the ADI cllectin f 82 dcuments and 35 queries which is available n-line in the SMART system. Clusters frmed are nt f unifrm size; ne r tw early clusters are exceptinally large. Variatin f the minimum crrelatin cutff value is an adequate cntrl fr the number f clusters generated. The effect f dcument er n the clustering methd is investigated, and the results are incnclusive. verall, the single-pass clustering methd is surprisingly effective and cmpares favrably with mre cmplicated clustering methds.. Intuctin In infrmatin retrieval, a given Item f infrmatin is ften repre- sented by an N-dimensinal vectr, each dimensin referring t a different prperty f the item. Using matrix algebra, cmparisns, classificatins, and ther prcesses can be perfrmed n the Infrmatin vectr r n a lng series f such vectrs. Classifying the item, that is, placing it in a grup f similar items, can save time later when it becmes necessary t refer t that Item again. Fr example, at the supermarket it is much easier t find the aisle with breakfast cereals, and then t lk fr the crn

2 XIV-2 flakes, than it is t search fr them n every shelf f every aisle. The classificatin prblem has tw basic aspects: ) classificatin already exists, and each new item is placed in that grup t which it is the mst similar; 2) n classificatin is assumed t exist and grups are frmed n the basis f similarities between items. [] Bth types f prcesses have advantages and disadvantages. tin already exists, the prcessing f new items is simple. If a classificawever, the size f the grupings ften becmes unbalanced in time with a resultant lss in efficiency. Merely finding the breakfast fd sectin wntt save much time if half the stre is stcked with cereals. The secnd classificatin prcedure f regruping every time a few new items are added t the set is time-cnsuming. Many cmparisns and decisins must be made t determine the ptimum grups. than ne grup. It may be reasnable t include the same item in mre But nce the time and truble have been taken t classify the items, any specific item is relatively easy t find. In the cntext f autmatic infrmatin retrieval, the items f infrmatin are dcuments and search requests, and the classificatin grups are called clusters. All the vectrs in a cluster are averaged t btain the cluster centrid, a vectr representatin f the entire grup. vectrs are crrelated with the centrids. Query nly thse centrids with suffi- ciently high crrelatins are expanded; that is each dcument in the cluster is crrelated with the query and retrieved in er f higher crrelatin. This type f search is knwn as a cluster search r tw-level serach. T cntrast, a full search wuld crrelate every dcument with the query and retrieve them in er f highest crrelatin. Cluster searching is a pr-

3 XIV-3 ven methd fr reducing search times hpefully withut significant lss f relevant dcuments. Many dcument gruping schemes representing bth slutins are in existence. Fr example, library catalging srts publicatins int already existing classificatins. n the ther hand, much research is being dne t implement the secnd classificatin prcess, that f frming clusters based nly n the similarities between the items. In general these methds require extensive cmputer time and strage space t calculate crrelatins between every pair f dcuments. RcchiTs and Bnner's clustering methds [2], [3] are f this kind, since they require a prcessing time prprtinal 2 t N, where N is the number f dcuments in the cllectin. A mre effi- cient clustering methd develped by Dattla [] des nt cmpute the dcument-dcument crrelatin matrix and reduces the prcessing time t ne f er N Lg N. A different, highly inexpensive methd f clustering has been prpsed; that f examining each dcument nly nce, and frming clusters in the prcess. Such a methd is called the single-pass clustering methd and is the subject f this paper. The single-pass clustering methd is appealing because it saves a significant amunt f time ver ther clustering methds. T illustrate, assume that a clustering methd requires minute f cmputer time fr dcuments. Then t cluster,0,0 dcuments, an N 2 clustering methd will take 70 years, an N Lg N methd 2 days, while the single-pass methd wuld take nly 7 days. The single-pass clustering methd als has a certain aesthetic appeal in that it represents a cmprmise between the tw basic aspects f the classificatin prblem. N initial classificatin exists, but as dcu-

4 XIV-4 ments are prcessed, classificatins are built up. A dcument is cnsidered fr inclusin int all existing clusters befre it is allwed t start a cluster f its wn. Classificatin grups are frmed principally n the basis f the similarities f an item with an existing grup rather than n the basis f similarities between items. wever, nce a dcument is accepted in an existing cluster, the cluster centrid is accingly revised. 2. The rgram A FRTRAN prgram was written t implement the single pass clustering methd. A flw diagram f the prcess is shwn in Fig.. The single-pass clustering methd assigns the first dcument vectr scanned as a cluster centrid. The next and succeeding dcuments are crrelated with existing cluster centrids. If a minimum cutff crrelatin is equaled r surpassed, the dcument is included in every such cluster and the cluster centrid is revised Cverlappingl, r the dcument is included nly in that cluster with the highest crrelatin, assuming the minimum cutff is again equaled r surpassed Cdisjint). If the minimum crrelatin is nt reached, a new cluster is frmed with the dcument as centrid. Tw crrelatin methds csine and Tanimt can be used with the prgram. Csine crrelatin nrmalizes all vectrs t length.0 in N-dimemsinal space, N being the number f cncept-dimensins in every dcument r query vectr* The csine f the angle between the tw vectrs is cmputed and used as a similarity measure. N Z.a 0 I d i d 2 Csrne S, = 2 = i=l N *",* 9 i=l N i=l 4 0

5 XIV-5 / TRL CARDS T INITIALIZE TYE F CRRELATIN CSINE, TANIMT, DISTANCE TYE F CLUSTERING: DISJINT, VERLAING MIN. CUTFF CRRELATIN, DC. RDER, ETC. READ ST DC READ NEXT DC > _ CRRELATE DC. WIT EXISTING CLUSTER CENTRIDS D ANY CRRELATINS EXCEED MIN. CUTFF CRRELATIN Y N 4 YES DISJINT/VERLAING DISJINT VERLAING INCLUDE DC. IN ALL CLUSTERS WIT CRR. ABVE MIN. CUTFF. UDATE CLUSTER CENTRIDS INCLUDE DC. IN CLUSTER WIT IGEST CRR. UDATE CLUSTER CENTRID DC. BECMES A CENTRID F A NEW CLUSTER -< < - N" AVE ALL DCS. BEEN RCESSED? Y YES RINT INFRMATIN FR CLUSTERS FRMED Y ABVE CLUSTERS - STRED IN SUITABLE FRMAT ARE FED INT SMART SEARC AND AVERAGE RGRAM T GET VARIUS MEASURES LIKE RECISIN-RECALL ETC. Q Single ass Clustering Methds Fig. Y

6 XIV-6 Tanimt T s crrelatin als perfrms vectr multiplicatin t btain a similarity measure. The denminatr is a nrmalizing term t prevent dense dcument vectrs frm being assigned disprprtinately high crrelatins. Tanimt S = " 2 d n d d.5 + d.d - d,.d l" " ~2"~2 ~"~2 N.. ~2 S d d 2 N.. N. _ N I (dv + I Cdi 2 - J i=l = = The prgram cntrl parameters are as fllws:. Type f clustering verlapping r disjint. 2. Crrelatin Methd csine r Tanimt. 3. Minimum Cncept Weight fr Centrid T prevent the dilutin f the cluster centrid, cncept weights f less than this value are set t zer. All runs were made at Cncept bund the highest cncept number f the dcument cllectin. 5. Mult Cncept weights f the nrmalized centrid vectr in the single-pass prgram are less than.0. T prevent mst f these cncepts frm disappearing when cnverting t integer (fixed pint) values fr the SMART system, all centrid weights are multiplied by MULT. All runs were made with MULT equal t., 6. Cutff The value f the minimum crrelatin cutff, belw which a dcument will nt be included in a cluster. Six cntrl cas cause the abve quantities t be intuced int the prgram, and als determine whether the utput will be printed r punched r

7 XIV-7 bth. Fur f the cas are name cas and place the prper SMART cntrl cas at the beginning f the utput t facilitate SMART evaluatin. 3. Investigatin and Results Using the 82 dcument, 35 query American Dcumentatin Institute cllectin filed n-line in the SMART S (Crnell Data Set), the fllwing investigatins f the single-pass clustering methd were undertaken: ) crrelatin cmparisn: csine crrelatin, frwa er cmparisn f clusters frmed by the tw different crrelatin, methds ver a range f crrelatin cutff values; 2) Disjint-verlapping cmparisn: csine crrelatin, frwa er cmparisn f disjint and verlapping clusters ver a range f crrelatin cutff values; 3). Variatin f dcument er: csine crrelatin, fixed crre- latin cutff cmparisn f bth verlapping and disjint clusters frmed with three different initial dcument ers; 4) SMART evaluatin: SMART evaluatin f the 6 sets f clusters frmed in part 3 t cmpare verlapping and disjint methds as well as the three dcument ers; al full search f the dcument cllectin and evaluatin fr cmparisn t single-pass runs; bl SMART evaluatin f clusters frmed by Dattla's methd, fr cmparisn; c) SMART evaluatin f Dattla!s clusters frmed frm initial single pass clusters. A) Crrelatin Cmparisn Fr csine crrelatins, crrelatin cutff values ranging frm.0 t.30 were investigated. The number f clusters frmed varied frm 5 at

8 XIV-8.0 t 3 at.30 with 4 clusters frmed at a crrelatin cutff f.20. Tanimt crrelatin cutff was varied frm.02 t.30. The number f clusters frmed varied frm 7 at.02 t 8 at.30 with 5 clusters frmed at a cutff f.04. All clustering was disjint using frwa dcument er. Appendix cntains tablesand graphs f the results f all runs perfrmed fr this phase f the investigatin. Fr every crrelatin methd, the relatinship between the number f clusters generated and the crrelatin cutff values is smth and nt unmanageably steep. Suitable variatin f the crrelatin cutff value can therefre be used t rughly cntrl the number f clusters that will be frmed fr a given dcument er. A striking fact abut csine and Tanimt clusters is the large variatin between sizes f clusters frmed, as shwn in Fig. 2. The largest clusters are usually amng the first 2 r 3 clusters, n dubt because they are frmed early and have a chance t inspect nearly the entire dcument cllectin. Tanimt clusters seem t be slightly mre evenly distributed than csine clusters. There is n apparent relatinship between dcuments included in a given csine cluster and dcuments included in a given Tanimt cluster. Bl Disjint-verlapping Cmparisn Fr csine crrelatin, frwa dcument er, the crrelatin cutff was varied frm.0 t.30, and bth verlapping and disjint clusters were frmed. were frmed. At a cutff f.20, 4 disjint and 7 verlapping clusters The average dcument is nly included in ne cluster fr the disjint methd but is included in 4.8 clusters fr the verlapping methd. Apparently the dilutin f the verlapping clusters resulted in the frma-

9 XIV-9 Csine.20 Crrelatin Cutff Tanimt.04 Crrelatin Cutff Number f Clusters Number f Dcuments in each Cluster Number f Clusters Number f dcuments in each Cluster 2 20 r ,6, r 7 3 8,9,0, r 2 3 r 4 5,6, r 7 6 r r 4 5 r 2 4 Ttal Clusters 5.9 Average Dcuments per Cluster 5 Ttal Clusters 5.5 Average Dcuments per Cluster Cluster Size Variatins Fig. 2

10 xiv-i tin f three mre clusters than the disjint methd. Any relatinship be- tween the dcument cmpsitin f a given verlap cluster and that f a gi/en disjint cluster is submerged by the large amunt f verlap. The results f the SMART evaluatin runs made n disjint and verlapping clusters will be discussed in that sectin. Tables and graphs fr all runs appear in Appendix. C) Variatin f Dcument er Three different initial dcument ers were used fr this investigatin frwa, reverse, and middle. The frwa dcument er lists the dcuments sequentially, while the reverse er has them backwas. Mid- dle er intuces the dcuments in an inside-ut er; fr example, dcuments, 2, 3, 4, 5, 6 in middle er wuld be 4, 3, 5, 2, 6,. Using a fixed csine crrelatin cutff f.20, bth disjint and verlapping clusters were frmed with the three different initial dcument ers. Fig. 3 shws the number f clusters frmed fr each situatin. Frm this data, it can be tentatively cncluded that a 0% r 20% variatin in the number f clusters frmed frm ne initial dcument er t anther will nt be unusual. D) SMART Evaluatin recisin and recall data were averaged ver all disjint and all verlapping runs t btain a measure f effectiveness reasnably independent f dcument er. recisin-recall graphs f the fllwing runs are pltted in Appendix 2.. Full search 2. Dattlafs cluster search

11 FL CFrwa verlap) - 7 clusters ML (Middle verlap) - RL (Reverse verlap) - FDJ (Frwa Disjint)- MDJ (Middle Disjint) - RDJ (Reverse Disjint)- Average Cluster Size Square Rt. f Std. DeV. 4 clusters 9 clusters 4 clusters 4 clusters 9 clusters 6.2 clusters f N 2 \ V2 =( X i ~ X AV j =2.07 clusters. V --i / Variatin in Cluster Size fr Different Dcument ers Fig. 3

12 XIV-2 3. Single-pass disjint cluster search 4. Single-pass verlapping cluster search The full search curve is naturally mre effective than ther curves. The single-pass disjint cluster search is the next best search. DattlaTs cluster search appears higher at bth ends f the graph than the singlepass verlapping cluster search, but lses sme grund in the middle t high-recall regin. Fig. 4 shws fr each run, the index f rank recall plus lg precisin, anther methd f system evaluatin, and reveals nearly the same results. Using ne set f disjint and ne set f verlapping clusters as initial clusters, DattlaTs clustering methd was evaluated. recisin- recall graphs f these results are fund in graph 2 f Appendix 2. The disjint initial clusters seem t give a higher precisin at lw recall values but a lwer precisin at high recall values than the verlapping initial clusters. Fr cmparisn DattlaTs previus run with unknwn initial clusters is als pltted and falls in the middle. In Appendix 2 graphs 3 and 4, the precisin-recall curves f the disjint and verlapping clusters used as initial clusters fr DattlaTs methd are pltted. Als Dattlafs results using these clusters as initial clusters are pltted t determine if any significant imprvement in clustering has ccurred. Fr the disjint clusters n imprvement ccurrs. Dattlafs clustering slightly reduces the effectiveness f the initial clusters. Fr the verlapping clusters, Dattlafs methd increases precisin at the lw recall end but reduces it at the high recall end f the graph, a significant imprvement. It is felt that the apparently high recall and precisin results b-

13 XIV-3 Search Type Rank Recall plus Lg recisin Full Average Disjint Cluster Average verlapping Cluster Dattla T s Cluster A arameter fr Cmparing Varius Methds Fig. 4

14 XIV-4 tained by the single-pass methd relative t DattlaTs methd is in part due t the expansin f ne r tw very large clusters, prviding a search f ver half the dcument cllectin. Fur mre runs were made t reduce the number f clusters expanded s that a mre legitimate cmparisn between the tw methds culd be made. Althugh a reductin in the number f clusters expanded ccurrs, n appreciable effect is bserved n the perfrmance measures, prbably due t the cntinued presence f thse extremely large clusters. Fr each run, the average dcuments checked per query are tabulated in Fig. 5. Mst single-pass runs checked far mre dcuments than DattlaTs runs, revealing that a single-pass cluster search is nt as efficient as a Dattla cluster search in terms f search time. Recall ceilings, hwever, naturally favr the system which lks at the mst dcuments. Glbal cmparisns f all evaluatin runs are tabulated in Table 2 f Appendix 2. Nrmalized precisin, nrmalized recall, rank recall, lg precisin, rank recall plus lg precisin, and recall ceiling are included. 4. Cnclusins ver the 82 dcument 35 query ADI cllectin, the single-pass clus- tering methd as described here is effective. Because the dcument cllec- tin is s small, these prmising results cannt be cnclusive. Mre ex- tensive evaluatin runs with larger dcument cllectins shuld be cnducted befre allwing the single-pass clustering methd a permanent place in the infrmatin retrieval wrld. The wide gap in the number f clusters frmed fr different input ers cntributes t a grwing suspicin that the single-pass methd may

15 XIV-5 be very er-dependent. The three ers used were nt truly randm. The middle er, fr example, is a cmpsite f bth the frwa and reverse ers as shwn in Fig. 6. Further investigatins f the single-pass clus- tering methd shuld include evaluatin with randm initial dcument ers t determine mre precisely the standa variatin f the number f clusters frmed at a given crrelatin cutff. Disjint clusters are mre effective than verlapping clusters. Re- ducing the figure f 4.3, the number f verlapping clusters t which the average dcument belngs, wuld certainly imprve th verlapping feature. It is recmmended that the verlapping algrithm be revised s that a dcument is included in the cluster with which it crrelates the highest (abve cutff) and nly in ne r tw ther clusters if the crrelatin is within the value f a small parameter, epsiln, f its highest crrelatin. The number f clusters which will be frmed can be effectively cntrlled by the establishment f a range f values fr the crrelatin cutff. erhaps the biggest pitfall f the single-pass methd is the frmatin f ne r tw excessively large clusters. It is these large clusters which cause t many dcuments t be checked and reduces the efficiency f the cluster search. T prevent large clusters frm frming, it Is recmmended that the single-pass clustering prgram be extended s that, as cluster size increases, the csine crrelatin cutff value slides up thrugh a sequence f values. As the number f items in a cluster increases, it becmes mre difficult fr a dcument t assimilate int that cluster. Such a prgram feature wuld reduce the large variatin in the sizes f the clusters. Finally the pssibility f mre than ne pass thrugh the dcument cllectin t determine the ptimum csine crrelatin cutff shuld be cn-

16 XIV-6 Run Descriptin Dattla (ML initial clusters) Dattla (MDJ initial clusters) RDJ (Reverse disjint) FDJ (Frwa disjint) MDJ (Middle disjint) ML (Middle verlapping) Average Dcuments Checked per Query Recall Ceiling Cmparisn f Average Dcuments fr Varius Methds Fig. 5 Frwa er Middle er Reverse er 3 2 Relatin in Three Dcument ers Taken An Illustratin Fig. 6

17 XIV-7 sidered. Fr large dcument cllectins a small representative subset culd be reprcessed until an ptimum crrelatin cutff is determined.

18 References [] Dattla, R. T.,A Fast Algrithm fr Autmatic Classificatin, Reprt ISR-4 t the Natinal Science Fundatin, Sectin V, Department f Cmputer Science, Crnell University, 968. {2] Bnner, R. E., "n Sme Clustering Techniques", IBM Jurnal f Research and Develpment, Vl. 8, N., January 964. [3] Rcchi, J. J., Jr., Dcument Retrieval Systems ptimizatin and Evaluatin, Reprt ISR-0 t the Natinal Science Fundatin, arva Cmputatin Labratry, Cambridge, Massachusetts, March 966.

19 XIV-9 Appendix Cluster Size Infrmatin

20 XIV-20 Cutff Number f Clusters Average Size f Clusters Maximum Size f Clusters Minimum Size f Clusters > >50.0 j Cluster Size bservatins fr Frwa Dcument er, Tanimt Crrelatin and Disjint Clusters Table

21 XIV-2 40 Dcument er: Frwa Crrelatin: Tanimt Cluster Type: Disjint 30 Number f Clusters *- 20 -w /Maximum Cluster Size 0 \ faverage Cluster Size 0 _L Cutff X Recall-recisin Graph Crrespnding t Table Graph

22 XIV-22 Cutff Number f Clusters Average Cluster Size Maximum Minimum Cluster Size bservatins fr Frwa Dcument er, Csine Crrelatin and Disjint Clusters Table 2

23 Dcument er: Frwa Crrelatin: Csine Cluster Type: Disjint Maximum Cluster Size Number f Clusters Cutff Recall-recisin Graph Crrespnding t Table 2 Graph 2

24 XV-24 Number f Clusters A Dcument Belngs T Average Maximum Minimum r-\ zf CM L [> C J- it CM cu TJ 4! c Q) 6 :3 0 Q u 4-J C 2 r b Cluster Size Number f Clusters Average Maximum Minimum r L [> I> L CM C CM zf L C zf zf C L r- [>- L r r r C 'Tj fc IU S & 0 U-. r 0 4 (/l c: 4 ' m > 0) LT) CI) N - C ci) +J (/I : a, rn ai r fc (D > T c nj c 0 +J ffl r-4 & CJ C C a C Cutff

25 XIV-25 Number f Clusters Disjint (frm Graph 2) CL Maximum Cluster Size AAverage Cluster Size A X ± Cutff Recall-recisin Graph Crrespnding t Table 3 Graph 3

26 XIV-26 ' ' - ' - - Number f Clusters A Dcument Belngs T Cluster Size Minimum ] Maximum Average Minimum Maximum Average Number f Clusters r r CM 5.9 -Zt" r r CM r C J- L 23.2 i r r CM C cn r C r L 23.5 en r! en CM en L r r 2.8 ' J- C 0 M m ) fc ',-. '/ C () a & IJ C C 4- J > & (D C.n cu M & a) 4 ' C 2 r C D CM ( i J 2 r ) r 4 J r () <M & a a :i r: x Dcument er Frwa Reverse Cluster Type Disjint verlapping Disjint verlapping Disjint 4-> a Middle verlapping

27 XIV-27 Appendix 2 Clustering Evaluatin

28 -28 Recall Full Search Dattla's Clusters Dattla s Clusters Using Initial Clus ter As ML MDJ 2 Sing le ass Clu sters Average* verlapping Average* Disjint ML = Single pass clusters with Csine Crrelatin, Cutff 0.2, Dcument er Middle, Cluster type verlapping. 2. MDJ = Same as abve, except Cluster type Disjint. * These Average recisins are taken ver three Dcument ers Recall-recisin Tables Table

29 XIV-29 Y A Full Search A Dattla's Cluster Search Single ass Disjint Cluster Search Single ass verlapping Cluster Search ( and are averages ver three dcument ers) 3 4 = 3h X \ K' QD- A-\ TI--D lh I I I I I I Recall Recall-recisin Graph Crrespnding t Table Graph

30 XIV-30 8 r~" attla T s Cluster Search Using Initial Clusters as Single ass Clusters (Disjint) Csine, 0.2 Cutff, Middle Dcument er Nt Using Single ass Clusters Using Initial Cluster as Single ass Clusters (verlap.) Csine, 0.2 Cutff, Recall * X Recall-recisin Graph Crrespnding t Table Graph 2

31 XIV-3.8 Y.7 A Single ass Clusters (Disjint) Middle Dcument er, Csine Crrelatin, Cutff Crrelatin 0.2 Dattla's Clusters (Disjint) Using Clusters in A as Initial Clusters A A Recall-recisin Graph Crrespnding t Table Graph 3

32 Y.8 A Single ass Clusters (verlapping) Dcument er Middle, Csine Crrelatin, Cutff 0.20 Dattla's Clusters (verlapping) Using Clusters in A as initial Clusters c (/) Recall Recall-recisin Graph Crrespnding t Table Graph 4

33 XIV-33 J w C L [> [> en ij cn C J- C C> a) N" I>. i 'U r. J i /"~N J S I **C J w k c -t c - I C cn en _r ) zf f c C xj CI) i: G JM I w 2 I> / -\ ~ J CM W 0) c. L C r r. t ) L - b > M, M", () <M -. I> "-"* b»- en d- Q L -f r >-> ' =t L t L. t «L x: n. -. / S -D Q L r S c>. IC) r L --J- L C. ' S *r L *"0 Q ;* 3 C) < - r- C C CJ re- > C if) w 0) T / -\ J" CM I r y <»"D Q S C L L C [> c > L r T) L c0 L C en L r 3 l> c cn N- d cu ci a p C - s. r C s J- 0>. t L C [> [ 4J [ Q C L L -t r C zr r-l lm C L C ) c :> p. < > / "> J ~ C / / c ' ' g3 C S / ~ ~ N r S N r r r M* C - C - e u CM a CD CM' b J r* C M M* ci) ' C 0 Q <l (/) -f- (0 -r C b U J n s~~< b b C -r < - > < -> 0 r, 'r ' r cd > I n p,. i n td, (D C (D - ' T I 0 - r ) ;: in 'i J - u U C r t p : cl). ' J n >; 0 J r-l :: C K () p d 'T j - ' d (/) (D, ' ] ; : C r-l 0 M i s_> p. r,m n :< j () r (0' (!). -n. (d (D C p r f C) N L ) C C ) : C 0) J.. p p (h 0!, r I C i' X C c 4 J (=u..ci) (0, In M hi) cn c > L L I> C v_> IX) [> [ J 0 C [.,(/') i j g Z C p. r - <> a i > p» «p I, L -, U C r i f ) p U U U U G J (0, \ ;* >.. C M r c. C b f j c d «C - LD Q t p ' ~ C L I <i) U -d L. M* r>. C ) C U r

A New Evaluation Measure. J. Joiner and L. Werner. The problems of evaluation and the needed criteria of evaluation

A New Evaluation Measure. J. Joiner and L. Werner. The problems of evaluation and the needed criteria of evaluation III-l III. A New Evaluatin Measure J. Jiner and L. Werner Abstract The prblems f evaluatin and the needed criteria f evaluatin measures in the SMART system f infrmatin retrieval are reviewed and discussed.

More information

The standards are taught in the following sequence.

The standards are taught in the following sequence. B L U E V A L L E Y D I S T R I C T C U R R I C U L U M MATHEMATICS Third Grade In grade 3, instructinal time shuld fcus n fur critical areas: (1) develping understanding f multiplicatin and divisin and

More information

ENSC Discrete Time Systems. Project Outline. Semester

ENSC Discrete Time Systems. Project Outline. Semester ENSC 49 - iscrete Time Systems Prject Outline Semester 006-1. Objectives The gal f the prject is t design a channel fading simulatr. Upn successful cmpletin f the prject, yu will reinfrce yur understanding

More information

A Matrix Representation of Panel Data

A Matrix Representation of Panel Data web Extensin 6 Appendix 6.A A Matrix Representatin f Panel Data Panel data mdels cme in tw brad varieties, distinct intercept DGPs and errr cmpnent DGPs. his appendix presents matrix algebra representatins

More information

5 th grade Common Core Standards

5 th grade Common Core Standards 5 th grade Cmmn Cre Standards In Grade 5, instructinal time shuld fcus n three critical areas: (1) develping fluency with additin and subtractin f fractins, and develping understanding f the multiplicatin

More information

CS 477/677 Analysis of Algorithms Fall 2007 Dr. George Bebis Course Project Due Date: 11/29/2007

CS 477/677 Analysis of Algorithms Fall 2007 Dr. George Bebis Course Project Due Date: 11/29/2007 CS 477/677 Analysis f Algrithms Fall 2007 Dr. Gerge Bebis Curse Prject Due Date: 11/29/2007 Part1: Cmparisn f Srting Algrithms (70% f the prject grade) The bjective f the first part f the assignment is

More information

How do scientists measure trees? What is DBH?

How do scientists measure trees? What is DBH? Hw d scientists measure trees? What is DBH? Purpse Students develp an understanding f tree size and hw scientists measure trees. Students bserve and measure tree ckies and explre the relatinship between

More information

Chapter 3: Cluster Analysis

Chapter 3: Cluster Analysis Chapter 3: Cluster Analysis } 3.1 Basic Cncepts f Clustering 3.1.1 Cluster Analysis 3.1. Clustering Categries } 3. Partitining Methds 3..1 The principle 3.. K-Means Methd 3..3 K-Medids Methd 3..4 CLARA

More information

Sequential Allocation with Minimal Switching

Sequential Allocation with Minimal Switching In Cmputing Science and Statistics 28 (1996), pp. 567 572 Sequential Allcatin with Minimal Switching Quentin F. Stut 1 Janis Hardwick 1 EECS Dept., University f Michigan Statistics Dept., Purdue University

More information

Lesson Plan. Recode: They will do a graphic organizer to sequence the steps of scientific method.

Lesson Plan. Recode: They will do a graphic organizer to sequence the steps of scientific method. Lessn Plan Reach: Ask the students if they ever ppped a bag f micrwave ppcrn and nticed hw many kernels were unppped at the bttm f the bag which made yu wnder if ther brands pp better than the ne yu are

More information

Sections 15.1 to 15.12, 16.1 and 16.2 of the textbook (Robbins-Miller) cover the materials required for this topic.

Sections 15.1 to 15.12, 16.1 and 16.2 of the textbook (Robbins-Miller) cover the materials required for this topic. Tpic : AC Fundamentals, Sinusidal Wavefrm, and Phasrs Sectins 5. t 5., 6. and 6. f the textbk (Rbbins-Miller) cver the materials required fr this tpic.. Wavefrms in electrical systems are current r vltage

More information

Subject description processes

Subject description processes Subject representatin 6.1.2. Subject descriptin prcesses Overview Fur majr prcesses r areas f practice fr representing subjects are classificatin, subject catalging, indexing, and abstracting. The prcesses

More information

Biplots in Practice MICHAEL GREENACRE. Professor of Statistics at the Pompeu Fabra University. Chapter 13 Offprint

Biplots in Practice MICHAEL GREENACRE. Professor of Statistics at the Pompeu Fabra University. Chapter 13 Offprint Biplts in Practice MICHAEL GREENACRE Prfessr f Statistics at the Pmpeu Fabra University Chapter 13 Offprint CASE STUDY BIOMEDICINE Cmparing Cancer Types Accrding t Gene Epressin Arrays First published:

More information

Hypothesis Tests for One Population Mean

Hypothesis Tests for One Population Mean Hypthesis Tests fr One Ppulatin Mean Chapter 9 Ala Abdelbaki Objective Objective: T estimate the value f ne ppulatin mean Inferential statistics using statistics in rder t estimate parameters We will be

More information

AP Statistics Notes Unit Two: The Normal Distributions

AP Statistics Notes Unit Two: The Normal Distributions AP Statistics Ntes Unit Tw: The Nrmal Distributins Syllabus Objectives: 1.5 The student will summarize distributins f data measuring the psitin using quartiles, percentiles, and standardized scres (z-scres).

More information

SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical model for microarray data analysis

SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical model for microarray data analysis SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical mdel fr micrarray data analysis David Rssell Department f Bistatistics M.D. Andersn Cancer Center, Hustn, TX 77030, USA rsselldavid@gmail.cm

More information

Dead-beat controller design

Dead-beat controller design J. Hetthéssy, A. Barta, R. Bars: Dead beat cntrller design Nvember, 4 Dead-beat cntrller design In sampled data cntrl systems the cntrller is realised by an intelligent device, typically by a PLC (Prgrammable

More information

Drought damaged area

Drought damaged area ESTIMATE OF THE AMOUNT OF GRAVEL CO~TENT IN THE SOIL BY A I R B O'RN EMS S D A T A Y. GOMI, H. YAMAMOTO, AND S. SATO ASIA AIR SURVEY CO., l d. KANAGAWA,JAPAN S.ISHIGURO HOKKAIDO TOKACHI UBPREFECTRAl OffICE

More information

Methods for Determination of Mean Speckle Size in Simulated Speckle Pattern

Methods for Determination of Mean Speckle Size in Simulated Speckle Pattern 0.478/msr-04-004 MEASUREMENT SCENCE REVEW, Vlume 4, N. 3, 04 Methds fr Determinatin f Mean Speckle Size in Simulated Speckle Pattern. Hamarvá, P. Šmíd, P. Hrváth, M. Hrabvský nstitute f Physics f the Academy

More information

Lead/Lag Compensator Frequency Domain Properties and Design Methods

Lead/Lag Compensator Frequency Domain Properties and Design Methods Lectures 6 and 7 Lead/Lag Cmpensatr Frequency Dmain Prperties and Design Methds Definitin Cnsider the cmpensatr (ie cntrller Fr, it is called a lag cmpensatr s K Fr s, it is called a lead cmpensatr Ntatin

More information

Math Foundations 20 Work Plan

Math Foundations 20 Work Plan Math Fundatins 20 Wrk Plan Units / Tpics 20.8 Demnstrate understanding f systems f linear inequalities in tw variables. Time Frame December 1-3 weeks 6-10 Majr Learning Indicatrs Identify situatins relevant

More information

Weathering. Title: Chemical and Mechanical Weathering. Grade Level: Subject/Content: Earth and Space Science

Weathering. Title: Chemical and Mechanical Weathering. Grade Level: Subject/Content: Earth and Space Science Weathering Title: Chemical and Mechanical Weathering Grade Level: 9-12 Subject/Cntent: Earth and Space Science Summary f Lessn: Students will test hw chemical and mechanical weathering can affect a rck

More information

Bootstrap Method > # Purpose: understand how bootstrap method works > obs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(obs) >

Bootstrap Method > # Purpose: understand how bootstrap method works > obs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(obs) > Btstrap Methd > # Purpse: understand hw btstrap methd wrks > bs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(bs) > mean(bs) [1] 21.64625 > # estimate f lambda > lambda = 1/mean(bs);

More information

Query Expansion. Lecture Objectives. Text Technologies for Data Science INFR Learn about Query Expansion. Implement: 10/24/2017

Query Expansion. Lecture Objectives. Text Technologies for Data Science INFR Learn about Query Expansion. Implement: 10/24/2017 Tet Technlgies fr Data Science INFR11145 Query Epansin Instructr: Walid Magdy 24-Oct-2017 Lecture Objectives Learn abut Query Epansin Query epansin methds Relevance feedback in IR Rcchi s algrithm PRF

More information

NUROP CONGRESS PAPER CHINESE PINYIN TO CHINESE CHARACTER CONVERSION

NUROP CONGRESS PAPER CHINESE PINYIN TO CHINESE CHARACTER CONVERSION NUROP Chinese Pinyin T Chinese Character Cnversin NUROP CONGRESS PAPER CHINESE PINYIN TO CHINESE CHARACTER CONVERSION CHIA LI SHI 1 AND LUA KIM TENG 2 Schl f Cmputing, Natinal University f Singapre 3 Science

More information

CAUSAL INFERENCE. Technical Track Session I. Phillippe Leite. The World Bank

CAUSAL INFERENCE. Technical Track Session I. Phillippe Leite. The World Bank CAUSAL INFERENCE Technical Track Sessin I Phillippe Leite The Wrld Bank These slides were develped by Christel Vermeersch and mdified by Phillippe Leite fr the purpse f this wrkshp Plicy questins are causal

More information

Resampling Methods. Chapter 5. Chapter 5 1 / 52

Resampling Methods. Chapter 5. Chapter 5 1 / 52 Resampling Methds Chapter 5 Chapter 5 1 / 52 1 51 Validatin set apprach 2 52 Crss validatin 3 53 Btstrap Chapter 5 2 / 52 Abut Resampling An imprtant statistical tl Pretending the data as ppulatin and

More information

CHAPTER 3 INEQUALITIES. Copyright -The Institute of Chartered Accountants of India

CHAPTER 3 INEQUALITIES. Copyright -The Institute of Chartered Accountants of India CHAPTER 3 INEQUALITIES Cpyright -The Institute f Chartered Accuntants f India INEQUALITIES LEARNING OBJECTIVES One f the widely used decisin making prblems, nwadays, is t decide n the ptimal mix f scarce

More information

Writing Guidelines. (Updated: November 25, 2009) Forwards

Writing Guidelines. (Updated: November 25, 2009) Forwards Writing Guidelines (Updated: Nvember 25, 2009) Frwards I have fund in my review f the manuscripts frm ur students and research assciates, as well as thse submitted t varius jurnals by thers that the majr

More information

CHAPTER 24: INFERENCE IN REGRESSION. Chapter 24: Make inferences about the population from which the sample data came.

CHAPTER 24: INFERENCE IN REGRESSION. Chapter 24: Make inferences about the population from which the sample data came. MATH 1342 Ch. 24 April 25 and 27, 2013 Page 1 f 5 CHAPTER 24: INFERENCE IN REGRESSION Chapters 4 and 5: Relatinships between tw quantitative variables. Be able t Make a graph (scatterplt) Summarize the

More information

4th Indian Institute of Astrophysics - PennState Astrostatistics School July, 2013 Vainu Bappu Observatory, Kavalur. Correlation and Regression

4th Indian Institute of Astrophysics - PennState Astrostatistics School July, 2013 Vainu Bappu Observatory, Kavalur. Correlation and Regression 4th Indian Institute f Astrphysics - PennState Astrstatistics Schl July, 2013 Vainu Bappu Observatry, Kavalur Crrelatin and Regressin Rahul Ry Indian Statistical Institute, Delhi. Crrelatin Cnsider a tw

More information

Revision: August 19, E Main Suite D Pullman, WA (509) Voice and Fax

Revision: August 19, E Main Suite D Pullman, WA (509) Voice and Fax .7.4: Direct frequency dmain circuit analysis Revisin: August 9, 00 5 E Main Suite D Pullman, WA 9963 (509) 334 6306 ice and Fax Overview n chapter.7., we determined the steadystate respnse f electrical

More information

Lab 1 The Scientific Method

Lab 1 The Scientific Method INTRODUCTION The fllwing labratry exercise is designed t give yu, the student, an pprtunity t explre unknwn systems, r universes, and hypthesize pssible rules which may gvern the behavir within them. Scientific

More information

CHAPTER Read Chapter 17, sections 1,2,3. End of Chapter problems: 25

CHAPTER Read Chapter 17, sections 1,2,3. End of Chapter problems: 25 CHAPTER 17 1. Read Chapter 17, sectins 1,2,3. End f Chapter prblems: 25 2. Suppse yu are playing a game that uses tw dice. If yu cunt the dts n the dice, yu culd have anywhere frm 2 t 12. The ways f prducing

More information

NUMBERS, MATHEMATICS AND EQUATIONS

NUMBERS, MATHEMATICS AND EQUATIONS AUSTRALIAN CURRICULUM PHYSICS GETTING STARTED WITH PHYSICS NUMBERS, MATHEMATICS AND EQUATIONS An integral part t the understanding f ur physical wrld is the use f mathematical mdels which can be used t

More information

Pattern Recognition 2014 Support Vector Machines

Pattern Recognition 2014 Support Vector Machines Pattern Recgnitin 2014 Supprt Vectr Machines Ad Feelders Universiteit Utrecht Ad Feelders ( Universiteit Utrecht ) Pattern Recgnitin 1 / 55 Overview 1 Separable Case 2 Kernel Functins 3 Allwing Errrs (Sft

More information

Least Squares Optimal Filtering with Multirate Observations

Least Squares Optimal Filtering with Multirate Observations Prc. 36th Asilmar Cnf. n Signals, Systems, and Cmputers, Pacific Grve, CA, Nvember 2002 Least Squares Optimal Filtering with Multirate Observatins Charles W. herrien and Anthny H. Hawes Department f Electrical

More information

the results to larger systems due to prop'erties of the projection algorithm. First, the number of hidden nodes must

the results to larger systems due to prop'erties of the projection algorithm. First, the number of hidden nodes must M.E. Aggune, M.J. Dambrg, M.A. El-Sharkawi, R.J. Marks II and L.E. Atlas, "Dynamic and static security assessment f pwer systems using artificial neural netwrks", Prceedings f the NSF Wrkshp n Applicatins

More information

MODULE FOUR. This module addresses functions. SC Academic Elementary Algebra Standards:

MODULE FOUR. This module addresses functions. SC Academic Elementary Algebra Standards: MODULE FOUR This mdule addresses functins SC Academic Standards: EA-3.1 Classify a relatinship as being either a functin r nt a functin when given data as a table, set f rdered pairs, r graph. EA-3.2 Use

More information

1 The limitations of Hartree Fock approximation

1 The limitations of Hartree Fock approximation Chapter: Pst-Hartree Fck Methds - I The limitatins f Hartree Fck apprximatin The n electrn single determinant Hartree Fck wave functin is the variatinal best amng all pssible n electrn single determinants

More information

WRITING THE REPORT. Organizing the report. Title Page. Table of Contents

WRITING THE REPORT. Organizing the report. Title Page. Table of Contents WRITING THE REPORT Organizing the reprt Mst reprts shuld be rganized in the fllwing manner. Smetime there is a valid reasn t include extra chapters in within the bdy f the reprt. 1. Title page 2. Executive

More information

MACE For Conformation Traits

MACE For Conformation Traits MACE Fr Cnfrmatin raits L. Klei and. J. Lawlr Hlstein Assciatin USA, Inc., Brattlebr, Vermnt, USA Intrductin Multiple acrss cuntry evaluatins (MACE) fr prductin traits are nw rutinely cmputed and used

More information

A Comparison of Methods for Computing the Eigenvalues and Eigenvectors of a Real Symmetric Matrix. By Paul A. White and Robert R.

A Comparison of Methods for Computing the Eigenvalues and Eigenvectors of a Real Symmetric Matrix. By Paul A. White and Robert R. A Cmparisn f Methds fr Cmputing the Eigenvalues and Eigenvectrs f a Real Symmetric Matrix By Paul A. White and Rbert R. Brwn Part I. The Eigenvalues I. Purpse. T cmpare thse methds fr cmputing the eigenvalues

More information

1996 Engineering Systems Design and Analysis Conference, Montpellier, France, July 1-4, 1996, Vol. 7, pp

1996 Engineering Systems Design and Analysis Conference, Montpellier, France, July 1-4, 1996, Vol. 7, pp THE POWER AND LIMIT OF NEURAL NETWORKS T. Y. Lin Department f Mathematics and Cmputer Science San Jse State University San Jse, Califrnia 959-003 tylin@cs.ssu.edu and Bereley Initiative in Sft Cmputing*

More information

DEFENSE OCCUPATIONAL AND ENVIRONMENTAL HEALTH READINESS SYSTEM (DOEHRS) ENVIRONMENTAL HEALTH SAMPLING ELECTRONIC DATA DELIVERABLE (EDD) GUIDE

DEFENSE OCCUPATIONAL AND ENVIRONMENTAL HEALTH READINESS SYSTEM (DOEHRS) ENVIRONMENTAL HEALTH SAMPLING ELECTRONIC DATA DELIVERABLE (EDD) GUIDE DEFENSE OCCUPATIOL AND ENVIRONMENTAL HEALTH READINESS SYSTEM (DOEHRS) ENVIRONMENTAL HEALTH SAMPLING ELECTRONIC DATA DELIVERABLE (EDD) GUIDE 20 JUNE 2017 V1.0 i TABLE OF CONTENTS 1 INTRODUCTION... 1 2 CONCEPT

More information

Part 3 Introduction to statistical classification techniques

Part 3 Introduction to statistical classification techniques Part 3 Intrductin t statistical classificatin techniques Machine Learning, Part 3, March 07 Fabi Rli Preamble ØIn Part we have seen that if we knw: Psterir prbabilities P(ω i / ) Or the equivalent terms

More information

THERMAL TEST LEVELS & DURATIONS

THERMAL TEST LEVELS & DURATIONS PREFERRED RELIABILITY PAGE 1 OF 7 PRACTICES PRACTICE NO. PT-TE-144 Practice: 1 Perfrm thermal dwell test n prtflight hardware ver the temperature range f +75 C/-2 C (applied at the thermal cntrl/munting

More information

Thermodynamics and Equilibrium

Thermodynamics and Equilibrium Thermdynamics and Equilibrium Thermdynamics Thermdynamics is the study f the relatinship between heat and ther frms f energy in a chemical r physical prcess. We intrduced the thermdynamic prperty f enthalpy,

More information

, which yields. where z1. and z2

, which yields. where z1. and z2 The Gaussian r Nrmal PDF, Page 1 The Gaussian r Nrmal Prbability Density Functin Authr: Jhn M Cimbala, Penn State University Latest revisin: 11 September 13 The Gaussian r Nrmal Prbability Density Functin

More information

Computational modeling techniques

Computational modeling techniques Cmputatinal mdeling techniques Lecture 4: Mdel checing fr ODE mdels In Petre Department f IT, Åb Aademi http://www.users.ab.fi/ipetre/cmpmd/ Cntent Stichimetric matrix Calculating the mass cnservatin relatins

More information

BASD HIGH SCHOOL FORMAL LAB REPORT

BASD HIGH SCHOOL FORMAL LAB REPORT BASD HIGH SCHOOL FORMAL LAB REPORT *WARNING: After an explanatin f what t include in each sectin, there is an example f hw the sectin might lk using a sample experiment Keep in mind, the sample lab used

More information

A study on GPS PDOP and its impact on position error

A study on GPS PDOP and its impact on position error IndianJurnalfRadi& SpacePhysics V1.26,April1997,pp. 107-111 A study n GPS and its impact n psitin errr P Banerjee,AnindyaBse& B SMathur TimeandFrequencySectin,NatinalPhysicalLabratry,NewDelhi110012 Received19June

More information

IN a recent article, Geary [1972] discussed the merit of taking first differences

IN a recent article, Geary [1972] discussed the merit of taking first differences The Efficiency f Taking First Differences in Regressin Analysis: A Nte J. A. TILLMAN IN a recent article, Geary [1972] discussed the merit f taking first differences t deal with the prblems that trends

More information

Floating Point Method for Solving Transportation. Problems with Additional Constraints

Floating Point Method for Solving Transportation. Problems with Additional Constraints Internatinal Mathematical Frum, Vl. 6, 20, n. 40, 983-992 Flating Pint Methd fr Slving Transprtatin Prblems with Additinal Cnstraints P. Pandian and D. Anuradha Department f Mathematics, Schl f Advanced

More information

A Regression Solution to the Problem of Criterion Score Comparability

A Regression Solution to the Problem of Criterion Score Comparability A Regressin Slutin t the Prblem f Criterin Scre Cmparability William M. Pugh Naval Health Research Center When the criterin measure in a study is the accumulatin f respnses r behavirs fr an individual

More information

ChE 471: LECTURE 4 Fall 2003

ChE 471: LECTURE 4 Fall 2003 ChE 47: LECTURE 4 Fall 003 IDEL RECTORS One f the key gals f chemical reactin engineering is t quantify the relatinship between prductin rate, reactr size, reactin kinetics and selected perating cnditins.

More information

CHAPTER 2 Algebraic Expressions and Fundamental Operations

CHAPTER 2 Algebraic Expressions and Fundamental Operations CHAPTER Algebraic Expressins and Fundamental Operatins OBJECTIVES: 1. Algebraic Expressins. Terms. Degree. Gruping 5. Additin 6. Subtractin 7. Multiplicatin 8. Divisin Algebraic Expressin An algebraic

More information

Phys. 344 Ch 7 Lecture 8 Fri., April. 10 th,

Phys. 344 Ch 7 Lecture 8 Fri., April. 10 th, Phys. 344 Ch 7 Lecture 8 Fri., April. 0 th, 009 Fri. 4/0 8. Ising Mdel f Ferrmagnets HW30 66, 74 Mn. 4/3 Review Sat. 4/8 3pm Exam 3 HW Mnday: Review fr est 3. See n-line practice test lecture-prep is t

More information

This is the most commonly used defuzzification technique. In this method, the overlapping area is counted twice.

This is the most commonly used defuzzification technique. In this method, the overlapping area is counted twice. Chapter 5 Defuzzificatin Methds Fuzzy rule based systems evaluate linguistic if-then rules using fuzzificatin, inference and cmpsitin prcedures. They prduce fuzzy results which usually have t be cnverted

More information

Determining the Accuracy of Modal Parameter Estimation Methods

Determining the Accuracy of Modal Parameter Estimation Methods Determining the Accuracy f Mdal Parameter Estimatin Methds by Michael Lee Ph.D., P.E. & Mar Richardsn Ph.D. Structural Measurement Systems Milpitas, CA Abstract The mst cmmn type f mdal testing system

More information

CHM112 Lab Graphing with Excel Grading Rubric

CHM112 Lab Graphing with Excel Grading Rubric Name CHM112 Lab Graphing with Excel Grading Rubric Criteria Pints pssible Pints earned Graphs crrectly pltted and adhere t all guidelines (including descriptive title, prperly frmatted axes, trendline

More information

Chemistry 20 Lesson 11 Electronegativity, Polarity and Shapes

Chemistry 20 Lesson 11 Electronegativity, Polarity and Shapes Chemistry 20 Lessn 11 Electrnegativity, Plarity and Shapes In ur previus wrk we learned why atms frm cvalent bnds and hw t draw the resulting rganizatin f atms. In this lessn we will learn (a) hw the cmbinatin

More information

Tree Structured Classifier

Tree Structured Classifier Tree Structured Classifier Reference: Classificatin and Regressin Trees by L. Breiman, J. H. Friedman, R. A. Olshen, and C. J. Stne, Chapman & Hall, 98. A Medical Eample (CART): Predict high risk patients

More information

COMP 551 Applied Machine Learning Lecture 5: Generative models for linear classification

COMP 551 Applied Machine Learning Lecture 5: Generative models for linear classification COMP 551 Applied Machine Learning Lecture 5: Generative mdels fr linear classificatin Instructr: Herke van Hf (herke.vanhf@mail.mcgill.ca) Slides mstly by: Jelle Pineau Class web page: www.cs.mcgill.ca/~hvanh2/cmp551

More information

Pipetting 101 Developed by BSU CityLab

Pipetting 101 Developed by BSU CityLab Discver the Micrbes Within: The Wlbachia Prject Pipetting 101 Develped by BSU CityLab Clr Cmparisns Pipetting Exercise #1 STUDENT OBJECTIVES Students will be able t: Chse the crrect size micrpipette fr

More information

GENERAL FORMULAS FOR FLAT-TOPPED WAVEFORMS. J.e. Sprott. Plasma Studies. University of Wisconsin

GENERAL FORMULAS FOR FLAT-TOPPED WAVEFORMS. J.e. Sprott. Plasma Studies. University of Wisconsin GENERAL FORMULAS FOR FLAT-TOPPED WAVEFORMS J.e. Sprtt PLP 924 September 1984 Plasma Studies University f Wiscnsin These PLP Reprts are infrmal and preliminary and as such may cntain errrs nt yet eliminated.

More information

MODULE 1. e x + c. [You can t separate a demominator, but you can divide a single denominator into each numerator term] a + b a(a + b)+1 = a + b

MODULE 1. e x + c. [You can t separate a demominator, but you can divide a single denominator into each numerator term] a + b a(a + b)+1 = a + b . REVIEW OF SOME BASIC ALGEBRA MODULE () Slving Equatins Yu shuld be able t slve fr x: a + b = c a d + e x + c and get x = e(ba +) b(c a) d(ba +) c Cmmn mistakes and strategies:. a b + c a b + a c, but

More information

The blessing of dimensionality for kernel methods

The blessing of dimensionality for kernel methods fr kernel methds Building classifiers in high dimensinal space Pierre Dupnt Pierre.Dupnt@ucluvain.be Classifiers define decisin surfaces in sme feature space where the data is either initially represented

More information

THE LIFE OF AN OBJECT IT SYSTEMS

THE LIFE OF AN OBJECT IT SYSTEMS THE LIFE OF AN OBJECT IT SYSTEMS Persns, bjects, r cncepts frm the real wrld, which we mdel as bjects in the IT system, have "lives". Actually, they have tw lives; the riginal in the real wrld has a life,

More information

Building to Transformations on Coordinate Axis Grade 5: Geometry Graph points on the coordinate plane to solve real-world and mathematical problems.

Building to Transformations on Coordinate Axis Grade 5: Geometry Graph points on the coordinate plane to solve real-world and mathematical problems. Building t Transfrmatins n Crdinate Axis Grade 5: Gemetry Graph pints n the crdinate plane t slve real-wrld and mathematical prblems. 5.G.1. Use a pair f perpendicular number lines, called axes, t define

More information

7 TH GRADE MATH STANDARDS

7 TH GRADE MATH STANDARDS ALGEBRA STANDARDS Gal 1: Students will use the language f algebra t explre, describe, represent, and analyze number expressins and relatins 7 TH GRADE MATH STANDARDS 7.M.1.1: (Cmprehensin) Select, use,

More information

Modelling of Clock Behaviour. Don Percival. Applied Physics Laboratory University of Washington Seattle, Washington, USA

Modelling of Clock Behaviour. Don Percival. Applied Physics Laboratory University of Washington Seattle, Washington, USA Mdelling f Clck Behaviur Dn Percival Applied Physics Labratry University f Washingtn Seattle, Washingtn, USA verheads and paper fr talk available at http://faculty.washingtn.edu/dbp/talks.html 1 Overview

More information

o o IMPORTANT REMINDERS Reports will be graded largely on their ability to clearly communicate results and important conclusions.

o o IMPORTANT REMINDERS Reports will be graded largely on their ability to clearly communicate results and important conclusions. BASD High Schl Frmal Lab Reprt GENERAL INFORMATION 12 pt Times New Rman fnt Duble-spaced, if required by yur teacher 1 inch margins n all sides (tp, bttm, left, and right) Always write in third persn (avid

More information

A New Approach to Increase Parallelism for Dependent Loops

A New Approach to Increase Parallelism for Dependent Loops A New Apprach t Increase Parallelism fr Dependent Lps Yeng-Sheng Chen, Department f Electrical Engineering, Natinal Taiwan University, Taipei 06, Taiwan, Tsang-Ming Jiang, Arctic Regin Supercmputing Center,

More information

Administrativia. Assignment 1 due thursday 9/23/2004 BEFORE midnight. Midterm exam 10/07/2003 in class. CS 460, Sessions 8-9 1

Administrativia. Assignment 1 due thursday 9/23/2004 BEFORE midnight. Midterm exam 10/07/2003 in class. CS 460, Sessions 8-9 1 Administrativia Assignment 1 due thursday 9/23/2004 BEFORE midnight Midterm eam 10/07/2003 in class CS 460, Sessins 8-9 1 Last time: search strategies Uninfrmed: Use nly infrmatin available in the prblem

More information

Differentiation Applications 1: Related Rates

Differentiation Applications 1: Related Rates Differentiatin Applicatins 1: Related Rates 151 Differentiatin Applicatins 1: Related Rates Mdel 1: Sliding Ladder 10 ladder y 10 ladder 10 ladder A 10 ft ladder is leaning against a wall when the bttm

More information

Department: MATHEMATICS

Department: MATHEMATICS Cde: MATH 022 Title: ALGEBRA SKILLS Institute: STEM Department: MATHEMATICS Curse Descriptin: This curse prvides students wh have cmpleted MATH 021 with the necessary skills and cncepts t cntinue the study

More information

Chem 115 POGIL Worksheet - Week 8 Thermochemistry (Continued), Electromagnetic Radiation, and Line Spectra

Chem 115 POGIL Worksheet - Week 8 Thermochemistry (Continued), Electromagnetic Radiation, and Line Spectra Chem 115 POGIL Wrksheet - Week 8 Thermchemistry (Cntinued), Electrmagnetic Radiatin, and Line Spectra Why? As we saw last week, enthalpy and internal energy are state functins, which means that the sum

More information

CHAPTER 4 DIAGNOSTICS FOR INFLUENTIAL OBSERVATIONS

CHAPTER 4 DIAGNOSTICS FOR INFLUENTIAL OBSERVATIONS CHAPTER 4 DIAGNOSTICS FOR INFLUENTIAL OBSERVATIONS 1 Influential bservatins are bservatins whse presence in the data can have a distrting effect n the parameter estimates and pssibly the entire analysis,

More information

A mathematical model for complete stress-strain curve prediction of permeable concrete

A mathematical model for complete stress-strain curve prediction of permeable concrete A mathematical mdel fr cmplete stress-strain curve predictin f permeable cncrete M. K. Hussin Y. Zhuge F. Bullen W. P. Lkuge Faculty f Engineering and Surveying, University f Suthern Queensland, Twmba,

More information

A Correlation of. to the. South Carolina Academic Standards for Mathematics Precalculus

A Correlation of. to the. South Carolina Academic Standards for Mathematics Precalculus A Crrelatin f Suth Carlina Academic Standards fr Mathematics Precalculus INTRODUCTION This dcument demnstrates hw Precalculus (Blitzer), 4 th Editin 010, meets the indicatrs f the. Crrelatin page references

More information

and the Doppler frequency rate f R , can be related to the coefficients of this polynomial. The relationships are:

and the Doppler frequency rate f R , can be related to the coefficients of this polynomial. The relationships are: Algrithm fr Estimating R and R - (David Sandwell, SIO, August 4, 2006) Azimith cmpressin invlves the alignment f successive eches t be fcused n a pint target Let s be the slw time alng the satellite track

More information

Admissibility Conditions and Asymptotic Behavior of Strongly Regular Graphs

Admissibility Conditions and Asymptotic Behavior of Strongly Regular Graphs Admissibility Cnditins and Asympttic Behavir f Strngly Regular Graphs VASCO MOÇO MANO Department f Mathematics University f Prt Oprt PORTUGAL vascmcman@gmailcm LUÍS ANTÓNIO DE ALMEIDA VIEIRA Department

More information

Medium Scale Integrated (MSI) devices [Sections 2.9 and 2.10]

Medium Scale Integrated (MSI) devices [Sections 2.9 and 2.10] EECS 270, Winter 2017, Lecture 3 Page 1 f 6 Medium Scale Integrated (MSI) devices [Sectins 2.9 and 2.10] As we ve seen, it s smetimes nt reasnable t d all the design wrk at the gate-level smetimes we just

More information

A computer model for input/output analysis in dynamic management information systems.

A computer model for input/output analysis in dynamic management information systems. Lehigh University Lehigh Preserve Theses and Dissertatins 1-1-1979 A cmputer mdel fr input/utput analysis in dynamic management infrmatin systems. Pefter W. Hartranft Fllw this and additinal wrks at: http://preserve.lehigh.edu/etd

More information

Experiment #3. Graphing with Excel

Experiment #3. Graphing with Excel Experiment #3. Graphing with Excel Study the "Graphing with Excel" instructins that have been prvided. Additinal help with learning t use Excel can be fund n several web sites, including http://www.ncsu.edu/labwrite/res/gt/gt-

More information

Technical Bulletin. Generation Interconnection Procedures. Revisions to Cluster 4, Phase 1 Study Methodology

Technical Bulletin. Generation Interconnection Procedures. Revisions to Cluster 4, Phase 1 Study Methodology Technical Bulletin Generatin Intercnnectin Prcedures Revisins t Cluster 4, Phase 1 Study Methdlgy Release Date: Octber 20, 2011 (Finalizatin f the Draft Technical Bulletin released n September 19, 2011)

More information

Copyright Paul Tobin 63

Copyright Paul Tobin 63 DT, Kevin t. lectric Circuit Thery DT87/ Tw-Prt netwrk parameters ummary We have seen previusly that a tw-prt netwrk has a pair f input terminals and a pair f utput terminals figure. These circuits were

More information

PSU GISPOPSCI June 2011 Ordinary Least Squares & Spatial Linear Regression in GeoDa

PSU GISPOPSCI June 2011 Ordinary Least Squares & Spatial Linear Regression in GeoDa There are tw parts t this lab. The first is intended t demnstrate hw t request and interpret the spatial diagnstics f a standard OLS regressin mdel using GeDa. The diagnstics prvide infrmatin abut the

More information

Cambridge Assessment International Education Cambridge Ordinary Level. Published

Cambridge Assessment International Education Cambridge Ordinary Level. Published Cambridge Assessment Internatinal Educatin Cambridge Ordinary Level ADDITIONAL MATHEMATICS 4037/1 Paper 1 Octber/Nvember 017 MARK SCHEME Maximum Mark: 80 Published This mark scheme is published as an aid

More information

Chapter 6 Fingerprints

Chapter 6 Fingerprints Chapter 6 Fingerprints Vcabulary: Arch: a fingerprint pattern in which the ridge pattern riginates frm ne side f the print and leaves frm the ther side Cre: Delta: a triangular ridge pattern with ridges

More information

Thermodynamics Partial Outline of Topics

Thermodynamics Partial Outline of Topics Thermdynamics Partial Outline f Tpics I. The secnd law f thermdynamics addresses the issue f spntaneity and invlves a functin called entrpy (S): If a prcess is spntaneus, then Suniverse > 0 (2 nd Law!)

More information

Interference is when two (or more) sets of waves meet and combine to produce a new pattern.

Interference is when two (or more) sets of waves meet and combine to produce a new pattern. Interference Interference is when tw (r mre) sets f waves meet and cmbine t prduce a new pattern. This pattern can vary depending n the riginal wave directin, wavelength, amplitude, etc. The tw mst extreme

More information

8.1. Review of experiments

8.1. Review of experiments VIII E V A L U A T I N VIII 1 E P E R I M E N T S Review f experiments As nted earlier,, the majr evaluatin tests f manual dexg systems which have been carried ut the last decade prvide the backgrund fr

More information

ECE 5318/6352 Antenna Engineering. Spring 2006 Dr. Stuart Long. Chapter 6. Part 7 Schelkunoff s Polynomial

ECE 5318/6352 Antenna Engineering. Spring 2006 Dr. Stuart Long. Chapter 6. Part 7 Schelkunoff s Polynomial ECE 538/635 Antenna Engineering Spring 006 Dr. Stuart Lng Chapter 6 Part 7 Schelkunff s Plynmial 7 Schelkunff s Plynmial Representatin (fr discrete arrays) AF( ψ ) N n 0 A n e jnψ N number f elements in

More information

ROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS

ROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS Particle Acceleratrs, 1986, Vl. 19, pp. 99-105 0031-2460/86/1904-0099/$15.00/0 1986 Grdn and Breach, Science Publishers, S.A. Printed in the United States f America ROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS

More information

COMP 551 Applied Machine Learning Lecture 9: Support Vector Machines (cont d)

COMP 551 Applied Machine Learning Lecture 9: Support Vector Machines (cont d) COMP 551 Applied Machine Learning Lecture 9: Supprt Vectr Machines (cnt d) Instructr: Herke van Hf (herke.vanhf@mail.mcgill.ca) Slides mstly by: Class web page: www.cs.mcgill.ca/~hvanh2/cmp551 Unless therwise

More information

MATCHING TECHNIQUES. Technical Track Session VI. Emanuela Galasso. The World Bank

MATCHING TECHNIQUES. Technical Track Session VI. Emanuela Galasso. The World Bank MATCHING TECHNIQUES Technical Track Sessin VI Emanuela Galass The Wrld Bank These slides were develped by Christel Vermeersch and mdified by Emanuela Galass fr the purpse f this wrkshp When can we use

More information