Analysis of Solar Generation and Weather Data in Smart Grid with Simultaneous Inference of Nonlinear Time Series

Size: px
Start display at page:

Download "Analysis of Solar Generation and Weather Data in Smart Grid with Simultaneous Inference of Nonlinear Time Series"

Transcription

1 Te First International Worksop on Smart Cities and Urban Informatics 215 Analysis of Solar Generation and Weater Data in Smart Grid wit Simultaneous Inference of Nonlinear Time Series Yu Wang, Guanqun Cao, Siwen Mao, and R. M. Nelms Department of Electrical and Computer Engineering, Auburn University, Auburn, AL Department of Matematics and Statistics, Auburn University, Auburn, AL Abstract Smart Grid is an important component of Smart City, were more renewable power generation and better energy management is required. Forecast on renewable power generation, from sources suc as solar and wind, is crucial for better energy management. However, te current forecast metods lack a compreensive understanding of te natural processes, and are tus limited in precise prediction. In tis paper, we introduce simultaneous inference to analyze te solar generation and weater data for better predictions. We first introduce a local linear model for nonlinear time series, and present te construction of te simultaneous confidence bands (SCB) of te time-varying coefficients, wic provide more information on te dynamic properties of te model. We ten use te simultaneous inference for solar intensity prediction using a real trace, were te superior performance of te proposed sceme is demonstrated over existing approaces. I. INTRODUCTION In recent years, as te development of te modern tecnologies in informatics, communication, control and computing, our living environment is becoming smart. Smart Home and Smart City ave gradually become part of our lives, and are no longer merely future concepts for te public. An important component of Smart City is te Smart Grid (SG), wic is regarded as te next generation power grid to create a widely distributed energy generation and delivery network. Te SG features te incorporation of power generation from renewable energy sources, especially solar and wind, wic meanwile requires a better energy management system in te SG [1]. Energy management in te SG as been studied in many previous works [1] [3]. It is indicated in [3] tat ig efficient power management cannot be realized witout a better forecast on te grid load and renewable power generation in te SG. Te problem of grid load forecasting as been studied by many researcers wit different tecniques suc as state space models [4], artificial neural networks and support vector macine (SVM) [5], and nonparametric functional time series analysis [3]. Te prediction on te renewable energy generation in te SG as also attracted great interest. Predictions on solar and wind generation can be found in [6] and [7] using SVM regression and JPDF forecast respectively. Because of te weater dependence nature of te forecasting problem, statistical metods can be found in almost every related literature. On te oter and, te wide range of applications elps te improvement of te statistical teory on nonparametric analysis [8] [1], and nonstationary time-series analysis [11], [12]. Recently, te autors of [13], [14] propose te metod of constructing a simultaneous confidence band (SCB) for time-varying coefficients. Tese researc advances improve te understanding of time-series, and provide better tecniques in renewable energy generation prediction. Te previous work on predicting solar power generation provides acceptable results using SVM regression [6]. However, by simply trying different SVM kernels after some basic data processing statistically, it lacks a deep analysis of te solar power generation and weater data, and tus is limited in precise predictions of oter data set. For example, te ceck of assumptions is missing on independence of variables and errors. Furtermore, te renewable energy generation is a function of weater variables, and is a stocastic process on nonlinear time series. Terefore, te relationsip between te weater variables and te power generation sould be analyzed over a long time for a compreensive understanding of teir dynamic relations. For example, a coefficient varying by time overall may stay constant for sort periods. Tese drawbacks will limit te applications for predictions in oter cases. Terefore, a metod tat can capture te dynamic property of te process would be igly desirable for better predictions on solar power generation in different scenarios. Motivated by tis observation, we introduce simultaneous inference of nonlinear time series proposed in [13] for understanding compreensively te deep and dynamic relationsip between renewable power generation process and te weater variable processes. Te simultaneous inference is based on te SCB of time-varying coefficients in te local linear model, wic we use for nonlinear time series analysis. It is based on te assumption of nonstationary processes for te error and weater variables, wic matces te case of our problem were many weater variables are sown to ave an obvious seasonal pattern, meaning te observations are not stationary assumed by many oter forecast tecniques. And te SCB sows te confidence bands of te coefficients over any lengt of time, wic can be used to test if te coefficients are truly time-varying or not. Tis elps us to refine te model by omitting variables wic tat are not significant. Te main contribution of tis paper is te introduction of /15/$ IEEE 6

2 Te First International Worksop on Smart Cities and Urban Informatics 215 te local linear model and SCB for analyzing te solar power generation as a nonlinear time series. By cecking te dynamic properties of te coefficients from its SCB, we are able to acieve a more compreensive understanding of te model, and based on tis, we can furter refine te model and use it for predicting te renewable energy generation. As an example, we apply it in predictions on daily solar power generation. Tis metod as a wide range of applications, wic is not limited to analyze and predict te solar energy generation. It can also be used for predictions in different time scales, from minutes to monts, depending on different purposes, and in oter cases, suc as wind power generation prediction. Te remainder of tis paper is organized as follows. We present te local linear model for nonlinear time-series analysis in Section II. Te construction of SCB for time-varying coefficients wit simulated results is introduced in Section III. We use te simultaneous inference for analyzing a trace of solar intensity and weater data in Section IV. Section V concludes tis paper. II. LOCAL LINEAR MODEL FOR NONLINEAR TIME SERIES We consider te power generation from a renewable source as a continuous-time stocastic process Y (t) and a function of te meteorological variables X(t), wic is a continuous-time covariate process. It follows tat Y (t) = f( X(t)), t R. To identify te function f( ), it is straigtforward to try a linear model first, as Y (t) = X T (t) β(t) + ɛ(t), t R, (1) were X(t) = (1, X1 (t),..., X p 1 (t)) T and β(t) = (β 1 (t),..., β p (t)) T are bot p 1 vectors, and ɛ(t) is te error at time t. To use tis model, we need to predict te regression coefficients β(t) at eac time t, so tat we can calculate Y (t) given te forecast on weater variables X(t). It is not difficult to obtain β(t) using parametric smooting metods suc as multiple linear regression. Altoug tis model can indicate a certain level of interactions between te variables X(t) and te response Y (t), it cannot be used to represent te real underlying process, especially wen te metod treat continuous-time series simply as discrete data points. As we will sow in Section IV-D, te prediction wit te linear model is not satisfactory, and cannot be used in practice. From te linear model, we notice tat te linearity is fairly strong between X(t) and Y (t) witin a sort time period, i.e., several days or a week. If we take advantage of tis property and consider te process as continuous-time, te model would be closer to te real process. For t i close to t, we can ave β(t i ) β(t) + (t i t) β (t), were β (t) is te derivative of β(t), and tus for any time t i close to t, we ave te local linear model as [15] Y (t i ) = X T (t i )( β(t) + (t i t) β (t)) + ɛ(t i ), t i t ±, (2) were te bandwidt is te size of te local neigborood. Tis model divides te time series into periods and creates linear models using local data. Tis way, we treat te data as a continuous-time series, and exploit te strong correlations between close time periods in weater dependent systems. A. Local Linear Estimation To identify te time-varying coefficients β(t), te least squares metod for linear regression can be used. We also add some weigts on te terms considering tat contributions from different neigbors are different, wic means a closer neigbor would ave a stronger effect, wile a furter neigbor weaker effect. Usually a kernel function K( ) is assigned to eac point, wic is a symmetric density function defined on [-1,1] [15]. Here, we use a popular Epanecnikov kernel. { 3(1 a K(a) = 2 )/4, if a 1, if a > 1, wic decays fast for remote data point. We ten ave te following weigted least squares problem to solve. ( ) min : (Y (t i ) X T (t i )( β(t) (t i t) β ti t (t)))k. t i t± (3) At eac time t, we solve for te coefficients ˆ β (t) and ˆ β (t) under te bandwidt. Suppose te total number of observations is n, we can pick t i simply as t i = i/n, 1 i n, and denote Y (t i ) as y i and X(t i ) as x i. From [9], we can solve (3) by calculating te following matrices S k (t) and R k (t): n ( ) k ( ) S k (t) = x i x T ti t ti t i K /(n) (4) i=1 n ( ) k ( ) ti t ti t R k (t) = x i y i K /(n), (5) i=1 were k =, 1, 2,... We ten ave ( ) ˆ β (t) ˆ = β (t) B. Selection of Bandwidt ( S (t) S T 1 (t) S 1 (t) S 2 (t) ) 1 ( R (t) R 1 (t) ). (6) To solve problem (3) for te complete model using (4) to (6), we need to first fix bandwidt. As discussed, is te bandwidt determining te size of data used to estimate for a local linear model at time t. If is too small, many useful points are not included for estimation, wic may cause a larger error in te model; if it is too large, more remote points are included, wic increases te computational complexity and reduces te smootness of te model. Terefore, it is important to coose a proper. Popular bandwidt selection tecniques can be found in [15], [16] for different applications. Te tecniques are different for constant bandwidt and variable bandwidt. For constant bandwidt selection considered in our model, we adopt te generalized cross-validation (GCV) tecnique [16], wic is suitable for a wide range of applications. Similar to multiple linear regression, te coefficients β are estimated from te observed data Y and X. Tus, a square 61

3 Te First International Worksop on Smart Cities and Urban Informatics 215 at matrix H() exists for ˆ Y = H() Y [17], depending on te bandwidt. Ten we can coose te bandwidt by { ˆ } Y Y 2 ĥ = arg min n(1 tr{h()}/n) 2, (7) were, tr( ) is te trace of te matrix, and n is te number of total observations. III. SIMULTANEOUS CONFIDENCE BAND FOR TIME-VARYING COEFFICIENTS In tis section, we introduce te basic conditions and construction of te simultaneous confidence band (SCB) metod proposed in [13], and ten discuss its implications to furter understand te modeling and predicting for te power generation process from te renewable energy sources based on te weater data. A. Model Assumptions and Asymptotic Normality Different from most current models for time series, te approac of SCB analysis assumes locally stationary processes for bot X(t) and ɛ(t) [12]. Te locally stationary process guarantees te stationary property for local time series, and is useful for local linear estimation. It actually belongs to a special class of non-stationary time series as x i = G(t i, F i ), ɛ i = H(t i, F i ), i = 1, 2,..., n, (8) were G(t i, F i ) and H(t i, F i ) are measurable functions well defined on t i [, 1], F i = (..., ξ i 1, ξ i ) wit {ξ i } i Z are independent and identically distributed (i.i.d.) random variables, and E(ɛ i x i ) =. In our model for renewable power generation processes, we furter assume tat {ɛ i } i Z are i.i.d. and dependent of {ξ i } i Z. Based on te above assumptions, te central limit teorem for ˆ β(t) states tat: supposing n and n 7 [13], ten for any fixed t (, 1), (n) 1/2 { ˆ β(t) β(t) 2 β (t)µ/2} N{, Σ 2 (t)}, (9) were µ = R x 2 K(x) dx, (1) Σ(t) = (M 1 (t)λ(t)m 1 (t)) 1/2, (11) M(t) = E( G(t, F ) G(t, F ) T ). (12) Te covariance matrix Λ(t) can be furter approximated using tecniques proposed in Section III-B. B. Simultaneous Confidence Band Deriving from te central limit property and basic assumptions sown above, te 1(1 α)% asymptotic simultaneous confidence tube of β C (t) can be constructed using te following formula: β C, (t) + ˆq 1 α ˆΣC (t)b s, (13) were β C, (t) is te bias corrected estimator defined in (14), B s = { z R s : z 1} is te unit ball, and s is te rank Algoritm 1: Construction of SCB for Time-varying Coefficients 1 Find a proper bandwidt ĥ from te GCV selector (7); 2 Let = 2ĥ and calculate β C, (t) using (14) and (3); 3 Obtain te estimated (1 α)t quantile ˆq 1 α via te bootstrap metod; 4 Estimate ˆM(t) = S (t ) and ˆΛ(t) by (??), and calculate ˆΣ C (t) according to (15); 5 Construct te 1(1 α)% SCB of β C (t) using (13). of a matrix C p s, wic we use for coosing different linear combinations of β(t), and β C (t) = C T β(t). To obtain te SCB, we simply take s = 1 in (13), and te SCB is constructed similarly to te confidence interval of te coefficients of te multiple linear regression: ˆβ ± tα/2,n p se( ˆβ), were se( ˆβ) is te standard error of ˆβ, and t α/2,n p is te upper α/2 percentage point of te t n 2 distribution [17]. Similarly, te first term is te estimator of te time-varying coefficients corrected for bias by β C, (t) = C T β (t) ( = C T 2 ) β / 2 (t) ˆ β (t), (14) were βĥ(t) can also be acquired by solving (3) using a corresponding kernel function K (a) = 2 2K( 2a) K(a) and an updated bandwidt = 2ĥ of te GCV selector ĥ. Te second term in (13) ˆq 1 α is actually te upper α/2 percentage point of te normal distribution N{, Σ 2 (t)} defined in (9), wile te tird term ˆΣ C (t) is te estimated standard error. Te metod of wild bootstrap is applied to obtain ˆq 1 α. Firstly, generate a large number i.i.d. vectors v 1, v 2,..., N(, I s ), were v i R p and I s denotes te s s identity matrix, and ten calculate q = sup t 1 n i=1 v ik ((t i t)/ )/(n ) ; repeat te previous step for a large number of times (say, 5) to acquire te estimated 1(1 α)% quantile ˆq 1 α of q. Te estimate of te standard error in (13), ˆΣ C (t) is defined similarly as (11). ˆΣ C (t) = (C T ˆM 1 (t)ˆλ(t) ˆM 1 (t)c) 1/2. (15) We sall estimate ˆM(t) and ˆΛ(t) respectively. From te definition of M(t) in (12), it can be estimated by ˆM(t) = S (t ), were S ( ) is defined in (4), and t = max{, min(t, 1 )}. To obtain ˆΛ(t), we first define two p 1 vectors Z i = x iˆɛ i and W i = m j= m Z i+j, a matrix Ω i = W iw T i /(2m + 1), and a function g(t, i) = K((t i t)/τ)/ n k=1 K((t k t), were m and τ can be simply cosen as m = n 2/7 and τ = n 1/7. Ten ˆΛ(t) can be calculated by ˆΛ(t) = n i=1 g(t, i)ω i. Tis way, we are able to calculate te SCB using all te estimates. Te above steps for constructing te SCB are summarized in Algoritm 1. C. Furter Discussions Te SCB provides a dynamic and compreensive view on β(t). In simple linear regression, te confidence interval 62

4 Te First International Worksop on Smart Cities and Urban Informatics 215 provides a measure of te overall quality of te regression line [17]. Similarly, te SCB illustrates te overall pattern of β(t) and tus te accuracy of te model. Confidence bands wit smaller widt implies a better model wit smaller variability, wile too wide confidence bands are limited in use. Note tat te SCB is constructed under a complete analysis on te continuous-time assumption, wic is not merely te connections of te point-wise confidence intervals on isolated time instances. Furter, te SCB can also be used to test weter te coefficients β(t) are truly time-varying or not. If a orizontal line is covered by te SCB of a β k (t), we accept te ypotesis tat β k (t) is constant and not time-varying. Furtermore, in different cases, we can construct te SCB for different linear combinations of β k (t) s by setting a different matrix C p s. For example, if we set C p 1 = [1, 1,,...], we obtain te SCB of β C (t) = C T β(t) = β 1 (t) + β 2 (t); if we set C p 2 = [1,,,...;, 1,,...], te SCB of β 1 (t) and β 2 (t) becomes a tube at any time t. Tis is because wen s = 2, te unit ball B 2 turns to a unit circle from a unit interval. Tis provides us a convenient way to furter test te model. D. Algoritm Performance for Simulated Processes We now simulate a model wit X(t) and ɛ(t) locally stationary processes discussed in Sec. III-A, and construct te 9% and 95% SCB respectively for a given model, to test te correctness by comparing wit te true results. We use te following local linear model wit time-varying coefficient: y i = β 1 (i/n) + β 2 (i/n)x i + ɛ i, (16) were β 1 (t) = cos(2πt)/4, and β 2 (t) = exp{ (t 1/2) 2 }/2. Define H(t i, F i ) = (1/2) j= a(t i) j ξ i j, G(ti, F i ) = (1; j= b(t)j ε i j ), were ξ k and ε l are i.i.d. N(, 1). Ten x i and ɛ i can be generated using (8), for i = 1, 2,..., n. For te above setting, we generate 5 samples of size 5, and for eac sample SCB is constructed wit bandwidts setting from.1 to.3 of step.25. We use 3 and 5 bootstrap samples to estimate ˆq 1 α for α =.1 and α =.5 to sow te effect of te sample size on te results. Te simulation results are sown in Table I, were te coverage rate and widt of SCB for β 2 (t) wit different bandwidts at 9% and 95% levels are listed. It sows tat te coverage rate is close to te nominal level wit most bandwidts. And te bandwidt selected by GCV is.22, wic yield fairly good results. We also notice tat ˆq 1 α are affected by te bootstrap sample size, and its value affects te widt of SCB directly. Terefore, for practical application, a large size of bootstrap samples is very important. According to our numerical studies, at least 5 samples are suggested. IV. APPLICATION TO SOLAR ENERGY GENERATION In tis section, we apply te SCB analysis to modeling te solar power generation process and predicting te generation amount based on weater data. We also compare results wit tat of oter metods. TABLE I THE COVERAGE PROBABILITIES OF SCB FOR β 2 (t) AND QUANTILES OF ˆq AT NOMINAL LEVEL OF 9% AND 95% Solar Intensity β 2 (t) 3 Samples 5 Samples bandwidt 9% 95% ˆq.9 ˆq.95 ˆq.9 ˆq A. Data Description Fig. 1. Daily solar intensity for 21 and 211. As an application, we consider te data from te UMASS Trace Repository [18], wic records te solar power generation by solar intensity in watts/m 2, and te data of several weater metrics from January, 21 to February, 213. It recorded te weater data every five minutes. Many weater parameters were observed in details. We consider five main variables, including temperature, umidity, dew point, wind speed, and precipitation. Te data as been used in [6] for a study of te statistical relationsip between te weater variables and solar power generation. Te paper also predicted solar power generation using multiple linear regression and Support Vector Macines regression. Our purpose is also to investigate te dynamic association between te weater variables and solar power generation, and to elp better predict te solar power generation. We plot te daily solar intensity for 21 and 211 in Fig. 1. An apparent seasonal pattern is sown wit peak in summer and valley in winter. It would be elpful to consider te seasonal patterns for forecast. It is also interesting to see a similar pattern for daily observations, and similar patterns can also be seen for te oter weater variables, suc as temperature, umidity, and dew point (See [6]). Fig. 1 also 63

5 Te First International Worksop on Smart Cities and Urban Informatics 215 sows a strong correlation between two consecutive days, wic means te solar generation process is not i.i.d. As discussed in Section III-A, we do not require te observations to be i.i.d. to construct te SCB. B. Prediction Model As in Section (II), we use te following local linear model. y i = β 1 (i/n) + 5 β p (i/n) x p,i + ɛ i, for i = 1,..., n, (17) p=2 were y i is te solar intensity, x p,i, p = 2, 3, 4, 5, represent te series of temperature in Fareneit, umidity in percentage, dew point in Fareneit, wind speed in miles per our, and precipitation in inces, respectively. We use n = 73 observations of 211 and 212 for local linear regression and te model is represented in a daily pattern. Note our model and analysis can be built on any time scale. We take daily pattern for an application example ere, were β 1 ( ) is te intercept and β p ( ) are te associated coefficients for x p,i. C. Simultaneous Inference for Time-varying Coefficients We now perform te SCB analysis. We center all te weater variables on teir averages so tat te intercept β 1 ( ) can be interpreted as te expected solar intensity. From GCV, we select te bandwidt =.25. Te 95% SCB of te coefficients β p ( ) are sown from Fig. 2 to 7. In eac figure, te middle tick solid curve is te estimated series for te variable; te upper and lower solid curves are te envelops for te simultaneous confidence band for eac variable. From te SCB, we are able to test weter a coefficients is significantly associated wit te solar intensity, wic equals to test: H : β p (t) =, t [, 1]; v.s. H 1 : β p (t), t [, 1]. If te zero line is included in te SCB, we accept te ypotesis tat te coefficient is not significant and could be omitted from te model; oterwise, we keep it in te model. We can also test weter te coefficients are constant, by attempting to include a constant orizontal line into te SCB. Tis is equal to testing: H : β p (t) = c p, t [, 1]; v.s. H 1 : β p (t) c p, t [, 1], were c p is a constant of eac p. If te line is covered, we accept tat te coefficient is constant; oterwise, it is not. In Fig. 2, te curve indicates te expected solar intensity for two years, and illustrates an obvious seasonal pattern. Te widt of te 95% SCB of β 1 (t) is so narrow tat no orizontal line can be covered, and even a iger level of 98% SCB cannot cover a orizontal line. We are confident tat te solar power generation is time-varying, te same as te natural process. As we center all te weater variables on teir averages, te SCB of te β p (t) actually indicates te effect on te solar intensity. In eac figure of Fig. 3 to 5, te zero line is not covered, wile in Fig. 6 and 7, te zero line is covered by te 95% SCB. Terefore, we can conclude tat for a level of 95%, temperature, umidity and dew point ave a TABLE II RMS-ERRORS IN watts/m 2 FOR TLLE, SVM AND MLR TLLE SVM MLR RMS-Error strong effect on solar generation, but te effect from wind and precipitation are weak. Also, we accept β 1 (t) to β 4 (t) as timevarying coefficients, because a constant orizontal line cannot be covered entirely in tose SCBs. Note te SCB associated wit wind in Fig. 6 sows some variations. Altoug te zero line may not be covered by a narrower SCB, say 9% SCB, at te 95% significant level, we do not accept β 5 (t) as a non-zero function.te SCB associated wit precipitation in Fig. 7 is also too wide for β 6 (t) to be accepted as a non-zero function. D. Comparisons wit Oter Models on Prediction Results From te above discussions, we could exclude te variable of precipitation and simplify te model for better prediction. We use te model to predict te daily solar intensity for January and February in 213. Te weater information of te previous year is used as te weater forecast, and te time for prediction is set from day 366 to 423, using data around January and February in 212. In oter words, te predictions are made using te data around te same time in te previous year. Te results are sown in Fig. 8. Te upper grap is te prediction curve made by te time-varying local linear model (TLLM) and te actual observations; te lower one sows te results from SVM regression used in [6]. We also perform te multiple linear regression (MLR) as in (1). But te prediction is too poor to be sown as a comparison ere. From Fig. 8, we can see tat te TLLE predicted curve tracks te actual observations better tan te SVM regression. And it is also sown in Table II tat te root mean squares error (RMS-Error) between te predicted series and te observations for TLLE is watts/m 2, smaller tan tat of SVM and MLR, wic are watts/m 2 and watts/m 2 respectively. Note tat te SVM regression depending igly on te selection of te parameters and te kernels, and tus, is not practical in many cases lacking a compreensive understanding of te real model. For example, te kernel function cosen for daily prediction is not guarantied to perform well in weekly prediction. However, te TLLE analyzes te model using simultaneous inference, wic reflects te overall pattern of te dynamic pattern of te regression functions. Terefore, it can be used in many oter applications, suc as te ourly or weekly solar power generation forecast were te time scale is set in our or week. V. CONCLUSIONS In tis paper, we proposed te simultaneous inference for weater dependent power generation from renewable energy sources, suc as solar and wind. We first introduced te local linear model for time series, and presented te construction of te SCB for time-varying coefficients. We ten performed te SCB analysis wit a trace of solar intensity and weater 64

6 Te First International Worksop on Smart Cities and Urban Informatics Intercept Temperature 2 1 Humidity Fig % SCB for Intercept Fig % SCB for Temperature Fig % SCB for Humidity x Dew Point Wind Percipitation Fig % SCB for Dew Point Fig % SCB for Wind Fig % SCB for Precipitation. Solar Intensity (watt\m2) Solar Intensity (watt\m2) Observed TLLE Observed SVM Fig. 8. Comparisons of predictions on solar intensity. data to validate te efficacy of te proposed approac. Te presented model was sown to outperform an existing metod for solar intensity prediction. ACKNOWLEDGMENT Tis work is supported in part by te US National Science Foundation under Grant CNS REFERENCES [1] X. Fang, S. Misra, G. Xue, and D. Yang, Smart grid te new and improved power grid: A survey, IEEE Commun. Surveys Tuts., vol. 14, no. 4, pp , Apr [2] Y. Wang, S. Mao, and R. Nelms, Distributed online algoritm for optimal real-time energy distribution in te smart grid, IEEE Internet Tings J., vol. 1, no. 1, pp. 7 8, Feb [3] M. Caouc, Clustering-based improvement of nonparametric functional time series forecasting: Application to intra-day ouseold-level load curves, IEEE Trans. Smart Grid, vol. 5, no. 1, pp , Jan [4] V. Dordonnat, S.J. Koopman, and M. Ooms, Dynamic factors in periodic time-varying regressions wit an application to ourly electricity load modelling, Comput. Stat. Data Anal., vol. 56, no. 11, pp , 212. [5] H.S. Hippert, C.E. Pedreira, and R.C. Souza, Neural networks for sortterm load forecasting: A review and evaluation, IEEE Trans. Power Syst., vol. 16, no. 1, pp , Feb. 21. [6] N. Sarma, P. Sarma, D. Irwin, and P. Senoy, Predicting solar generation from weater forecasts using macine learning, in Proc. IEEE SmartGridComm 11, Oct. 211, pp [7] S. Zu, M. Yang, M. Liu, and W.J. Lee, One parametric approac for sort-term jpdf forecast of wind generation, in Proc. 213 IEEE IAS Annual Meeting, Orlando, FL, Oct. 213, pp [8] F. Ferraty and P. Vieu, Nonparametric Functional Data Analysis: Teory and Practice. New York: Springer-Verlag, 26. [9] W. Hardle, Applied nonparametric regression, Econometric Teory, vol. 8, no. 3, pp , Sept [1] G. Cao, L. Yang, and D. Todem, Simultaneous inference for te mean function based on dense functional data, J. Nonparametric Stat., vol. 24, no. 2, pp , June 212. [11] Z. Zou and W.B. Wu, Local linear quantile estimation of nonstationary time series, Ann. Statist., vol. 37, no. 5B, pp , July 29. [12] D. Dragicescu, S. Guillas, and W.B. Wu, Quantile curve estimation and visualization for nonstationary time series, J. Comput. Grap. Statist., vol. 18, no. 1, pp. 1 2, 29. [13] Z. Zou and W.B. Wu, Simultaneous inference of linear models wit time varying coefficients, Journal of te Royal Statistical Society: Series B (Statistical Metodology), vol. 72, no. 4, pp , Sept. 21. [14] G. Cao, Simultaneous confidence bands for derivatives of dependent functional data, Elec. J. Stat., vol. 8, no. 2, pp , Dec [15] J. Fan and I. Gijbels, Local Polynomial Modelling and Its Applications. New York: Capman and Hall, [16] P. Craven and G. Waba, Smooting noisy data wit spline functions, Numer. Mat, vol. 31, no. 4, pp , Dec [17] E. Montgomery, C.and Peck and G. Vining, Introduction to Linear Regression Analysis, 4t ed. A Jon Wiley & Sons, Inc., July [18] [online] Available: ttp://traces.cs.umass.edu/. 65

The Priestley-Chao Estimator

The Priestley-Chao Estimator Te Priestley-Cao Estimator In tis section we will consider te Pristley-Cao estimator of te unknown regression function. It is assumed tat we ave a sample of observations (Y i, x i ), i = 1,..., n wic are

More information

A Jump-Preserving Curve Fitting Procedure Based On Local Piecewise-Linear Kernel Estimation

A Jump-Preserving Curve Fitting Procedure Based On Local Piecewise-Linear Kernel Estimation A Jump-Preserving Curve Fitting Procedure Based On Local Piecewise-Linear Kernel Estimation Peiua Qiu Scool of Statistics University of Minnesota 313 Ford Hall 224 Curc St SE Minneapolis, MN 55455 Abstract

More information

Adaptive Learning Hybrid Model for Solar Intensity Forecasting

Adaptive Learning Hybrid Model for Solar Intensity Forecasting The final version of record is available at http://dx.doi.org/1.119/tii.217.2789289 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL.XXX, NO.XXX, MONTH YEAR 1 Adaptive Learning Hybrid Model for Solar Intensity

More information

Basic Nonparametric Estimation Spring 2002

Basic Nonparametric Estimation Spring 2002 Basic Nonparametric Estimation Spring 2002 Te following topics are covered today: Basic Nonparametric Regression. Tere are four books tat you can find reference: Silverman986, Wand and Jones995, Hardle990,

More information

Kernel Density Based Linear Regression Estimate

Kernel Density Based Linear Regression Estimate Kernel Density Based Linear Regression Estimate Weixin Yao and Zibiao Zao Abstract For linear regression models wit non-normally distributed errors, te least squares estimate (LSE will lose some efficiency

More information

Bandwidth Selection in Nonparametric Kernel Testing

Bandwidth Selection in Nonparametric Kernel Testing Te University of Adelaide Scool of Economics Researc Paper No. 2009-0 January 2009 Bandwidt Selection in Nonparametric ernel Testing Jiti Gao and Irene Gijbels Bandwidt Selection in Nonparametric ernel

More information

EFFICIENCY OF MODEL-ASSISTED REGRESSION ESTIMATORS IN SAMPLE SURVEYS

EFFICIENCY OF MODEL-ASSISTED REGRESSION ESTIMATORS IN SAMPLE SURVEYS Statistica Sinica 24 2014, 395-414 doi:ttp://dx.doi.org/10.5705/ss.2012.064 EFFICIENCY OF MODEL-ASSISTED REGRESSION ESTIMATORS IN SAMPLE SURVEYS Jun Sao 1,2 and Seng Wang 3 1 East Cina Normal University,

More information

Financial Econometrics Prof. Massimo Guidolin

Financial Econometrics Prof. Massimo Guidolin CLEFIN A.A. 2010/2011 Financial Econometrics Prof. Massimo Guidolin A Quick Review of Basic Estimation Metods 1. Were te OLS World Ends... Consider two time series 1: = { 1 2 } and 1: = { 1 2 }. At tis

More information

Bootstrap confidence intervals in nonparametric regression without an additive model

Bootstrap confidence intervals in nonparametric regression without an additive model Bootstrap confidence intervals in nonparametric regression witout an additive model Dimitris N. Politis Abstract Te problem of confidence interval construction in nonparametric regression via te bootstrap

More information

A New Diagnostic Test for Cross Section Independence in Nonparametric Panel Data Model

A New Diagnostic Test for Cross Section Independence in Nonparametric Panel Data Model e University of Adelaide Scool of Economics Researc Paper No. 2009-6 October 2009 A New Diagnostic est for Cross Section Independence in Nonparametric Panel Data Model Jia Cen, Jiti Gao and Degui Li e

More information

On Local Linear Regression Estimation of Finite Population Totals in Model Based Surveys

On Local Linear Regression Estimation of Finite Population Totals in Model Based Surveys American Journal of Teoretical and Applied Statistics 2018; 7(3): 92-101 ttp://www.sciencepublisinggroup.com/j/ajtas doi: 10.11648/j.ajtas.20180703.11 ISSN: 2326-8999 (Print); ISSN: 2326-9006 (Online)

More information

Model Specification Testing in Nonparametric and Semiparametric Time Series Econometrics 1

Model Specification Testing in Nonparametric and Semiparametric Time Series Econometrics 1 Model Specification Testing in Nonparametric and Semiparametric Time Series Econometrics 1 By Jiti Gao 2 and Maxwell King 3 Abstract We propose a simultaneous model specification procedure for te conditional

More information

Fast optimal bandwidth selection for kernel density estimation

Fast optimal bandwidth selection for kernel density estimation Fast optimal bandwidt selection for kernel density estimation Vikas Candrakant Raykar and Ramani Duraiswami Dept of computer science and UMIACS, University of Maryland, CollegePark {vikas,ramani}@csumdedu

More information

Chapter 1. Density Estimation

Chapter 1. Density Estimation Capter 1 Density Estimation Let X 1, X,..., X n be observations from a density f X x. Te aim is to use only tis data to obtain an estimate ˆf X x of f X x. Properties of f f X x x, Parametric metods f

More information

A = h w (1) Error Analysis Physics 141

A = h w (1) Error Analysis Physics 141 Introduction In all brances of pysical science and engineering one deals constantly wit numbers wic results more or less directly from experimental observations. Experimental observations always ave inaccuracies.

More information

Fast Exact Univariate Kernel Density Estimation

Fast Exact Univariate Kernel Density Estimation Fast Exact Univariate Kernel Density Estimation David P. Hofmeyr Department of Statistics and Actuarial Science, Stellenbosc University arxiv:1806.00690v2 [stat.co] 12 Jul 2018 July 13, 2018 Abstract Tis

More information

Bootstrap prediction intervals for Markov processes

Bootstrap prediction intervals for Markov processes arxiv: arxiv:0000.0000 Bootstrap prediction intervals for Markov processes Li Pan and Dimitris N. Politis Li Pan Department of Matematics University of California San Diego La Jolla, CA 92093-0112, USA

More information

232 Calculus and Structures

232 Calculus and Structures 3 Calculus and Structures CHAPTER 17 JUSTIFICATION OF THE AREA AND SLOPE METHODS FOR EVALUATING BEAMS Calculus and Structures 33 Copyrigt Capter 17 JUSTIFICATION OF THE AREA AND SLOPE METHODS 17.1 THE

More information

Parameter Fitted Scheme for Singularly Perturbed Delay Differential Equations

Parameter Fitted Scheme for Singularly Perturbed Delay Differential Equations International Journal of Applied Science and Engineering 2013. 11, 4: 361-373 Parameter Fitted Sceme for Singularly Perturbed Delay Differential Equations Awoke Andargiea* and Y. N. Reddyb a b Department

More information

HOW TO DEAL WITH FFT SAMPLING INFLUENCES ON ADEV CALCULATIONS

HOW TO DEAL WITH FFT SAMPLING INFLUENCES ON ADEV CALCULATIONS HOW TO DEAL WITH FFT SAMPLING INFLUENCES ON ADEV CALCULATIONS Po-Ceng Cang National Standard Time & Frequency Lab., TL, Taiwan 1, Lane 551, Min-Tsu Road, Sec. 5, Yang-Mei, Taoyuan, Taiwan 36 Tel: 886 3

More information

A MONTE CARLO ANALYSIS OF THE EFFECTS OF COVARIANCE ON PROPAGATED UNCERTAINTIES

A MONTE CARLO ANALYSIS OF THE EFFECTS OF COVARIANCE ON PROPAGATED UNCERTAINTIES A MONTE CARLO ANALYSIS OF THE EFFECTS OF COVARIANCE ON PROPAGATED UNCERTAINTIES Ronald Ainswort Hart Scientific, American Fork UT, USA ABSTRACT Reports of calibration typically provide total combined uncertainties

More information

Lecture 15. Interpolation II. 2 Piecewise polynomial interpolation Hermite splines

Lecture 15. Interpolation II. 2 Piecewise polynomial interpolation Hermite splines Lecture 5 Interpolation II Introduction In te previous lecture we focused primarily on polynomial interpolation of a set of n points. A difficulty we observed is tat wen n is large, our polynomial as to

More information

1 The concept of limits (p.217 p.229, p.242 p.249, p.255 p.256) 1.1 Limits Consider the function determined by the formula 3. x since at this point

1 The concept of limits (p.217 p.229, p.242 p.249, p.255 p.256) 1.1 Limits Consider the function determined by the formula 3. x since at this point MA00 Capter 6 Calculus and Basic Linear Algebra I Limits, Continuity and Differentiability Te concept of its (p.7 p.9, p.4 p.49, p.55 p.56). Limits Consider te function determined by te formula f Note

More information

lecture 26: Richardson extrapolation

lecture 26: Richardson extrapolation 43 lecture 26: Ricardson extrapolation 35 Ricardson extrapolation, Romberg integration Trougout numerical analysis, one encounters procedures tat apply some simple approximation (eg, linear interpolation)

More information

Notes on Neural Networks

Notes on Neural Networks Artificial neurons otes on eural etwors Paulo Eduardo Rauber 205 Consider te data set D {(x i y i ) i { n} x i R m y i R d } Te tas of supervised learning consists on finding a function f : R m R d tat

More information

Artificial Neural Network Model Based Estimation of Finite Population Total

Artificial Neural Network Model Based Estimation of Finite Population Total International Journal of Science and Researc (IJSR), India Online ISSN: 2319-7064 Artificial Neural Network Model Based Estimation of Finite Population Total Robert Kasisi 1, Romanus O. Odiambo 2, Antony

More information

Continuous Stochastic Processes

Continuous Stochastic Processes Continuous Stocastic Processes Te term stocastic is often applied to penomena tat vary in time, wile te word random is reserved for penomena tat vary in space. Apart from tis distinction, te modelling

More information

How to Find the Derivative of a Function: Calculus 1

How to Find the Derivative of a Function: Calculus 1 Introduction How to Find te Derivative of a Function: Calculus 1 Calculus is not an easy matematics course Te fact tat you ave enrolled in suc a difficult subject indicates tat you are interested in te

More information

Online Appendix to Estimating Fiscal Limits: The Case of Greece

Online Appendix to Estimating Fiscal Limits: The Case of Greece Online Appendix to Estimating Fiscal Limits: Te Case of Greece Huixin Bi and Nora Traum May 5, Tis appendix includes details of nonlinear numerical solutions, estimation diagnostics and additional results

More information

Department of Econometrics and Business Statistics

Department of Econometrics and Business Statistics ISSN 1440-771X Australia Department of Econometrics and Business Statistics ttp://www.buseco.monas.edu.au/depts/ebs/pubs/wpapers/ Bayesian Bandwidt Estimation in Nonparametric Time-Varying Coefficient

More information

Poisson Equation in Sobolev Spaces

Poisson Equation in Sobolev Spaces Poisson Equation in Sobolev Spaces OcMountain Dayligt Time. 6, 011 Today we discuss te Poisson equation in Sobolev spaces. It s existence, uniqueness, and regularity. Weak Solution. u = f in, u = g on

More information

THE STURM-LIOUVILLE-TRANSFORMATION FOR THE SOLUTION OF VECTOR PARTIAL DIFFERENTIAL EQUATIONS. L. Trautmann, R. Rabenstein

THE STURM-LIOUVILLE-TRANSFORMATION FOR THE SOLUTION OF VECTOR PARTIAL DIFFERENTIAL EQUATIONS. L. Trautmann, R. Rabenstein Worksop on Transforms and Filter Banks (WTFB),Brandenburg, Germany, Marc 999 THE STURM-LIOUVILLE-TRANSFORMATION FOR THE SOLUTION OF VECTOR PARTIAL DIFFERENTIAL EQUATIONS L. Trautmann, R. Rabenstein Lerstul

More information

Uniform Consistency for Nonparametric Estimators in Null Recurrent Time Series

Uniform Consistency for Nonparametric Estimators in Null Recurrent Time Series Te University of Adelaide Scool of Economics Researc Paper No. 2009-26 Uniform Consistency for Nonparametric Estimators in Null Recurrent Time Series Jiti Gao, Degui Li and Dag Tjøsteim Te University of

More information

4. The slope of the line 2x 7y = 8 is (a) 2/7 (b) 7/2 (c) 2 (d) 2/7 (e) None of these.

4. The slope of the line 2x 7y = 8 is (a) 2/7 (b) 7/2 (c) 2 (d) 2/7 (e) None of these. Mat 11. Test Form N Fall 016 Name. Instructions. Te first eleven problems are wort points eac. Te last six problems are wort 5 points eac. For te last six problems, you must use relevant metods of algebra

More information

7 Semiparametric Methods and Partially Linear Regression

7 Semiparametric Methods and Partially Linear Regression 7 Semiparametric Metods and Partially Linear Regression 7. Overview A model is called semiparametric if it is described by and were is nite-dimensional (e.g. parametric) and is in nite-dimensional (nonparametric).

More information

NADARAYA WATSON ESTIMATE JAN 10, 2006: version 2. Y ik ( x i

NADARAYA WATSON ESTIMATE JAN 10, 2006: version 2. Y ik ( x i NADARAYA WATSON ESTIMATE JAN 0, 2006: version 2 DATA: (x i, Y i, i =,..., n. ESTIMATE E(Y x = m(x by n i= ˆm (x = Y ik ( x i x n i= K ( x i x EXAMPLES OF K: K(u = I{ u c} (uniform or box kernel K(u = u

More information

Math 102 TEST CHAPTERS 3 & 4 Solutions & Comments Fall 2006

Math 102 TEST CHAPTERS 3 & 4 Solutions & Comments Fall 2006 Mat 102 TEST CHAPTERS 3 & 4 Solutions & Comments Fall 2006 f(x+) f(x) 10 1. For f(x) = x 2 + 2x 5, find ))))))))) and simplify completely. NOTE: **f(x+) is NOT f(x)+! f(x+) f(x) (x+) 2 + 2(x+) 5 ( x 2

More information

Kernel Smoothing and Tolerance Intervals for Hierarchical Data

Kernel Smoothing and Tolerance Intervals for Hierarchical Data Clemson University TigerPrints All Dissertations Dissertations 12-2016 Kernel Smooting and Tolerance Intervals for Hierarcical Data Cristoper Wilson Clemson University, cwilso6@clemson.edu Follow tis and

More information

Long Term Time Series Prediction with Multi-Input Multi-Output Local Learning

Long Term Time Series Prediction with Multi-Input Multi-Output Local Learning Long Term Time Series Prediction wit Multi-Input Multi-Output Local Learning Gianluca Bontempi Macine Learning Group, Département d Informatique Faculté des Sciences, ULB, Université Libre de Bruxelles

More information

Online Learning: Bandit Setting

Online Learning: Bandit Setting Online Learning: Bandit Setting Daniel asabi Summer 04 Last Update: October 0, 06 Introduction [TODO Bandits. Stocastic setting Suppose tere exists unknown distributions ν,..., ν, suc tat te loss at eac

More information

WYSE Academic Challenge 2004 Sectional Mathematics Solution Set

WYSE Academic Challenge 2004 Sectional Mathematics Solution Set WYSE Academic Callenge 00 Sectional Matematics Solution Set. Answer: B. Since te equation can be written in te form x + y, we ave a major 5 semi-axis of lengt 5 and minor semi-axis of lengt. Tis means

More information

REVIEW LAB ANSWER KEY

REVIEW LAB ANSWER KEY REVIEW LAB ANSWER KEY. Witout using SN, find te derivative of eac of te following (you do not need to simplify your answers): a. f x 3x 3 5x x 6 f x 3 3x 5 x 0 b. g x 4 x x x notice te trick ere! x x g

More information

1. Questions (a) through (e) refer to the graph of the function f given below. (A) 0 (B) 1 (C) 2 (D) 4 (E) does not exist

1. Questions (a) through (e) refer to the graph of the function f given below. (A) 0 (B) 1 (C) 2 (D) 4 (E) does not exist Mat 1120 Calculus Test 2. October 18, 2001 Your name Te multiple coice problems count 4 points eac. In te multiple coice section, circle te correct coice (or coices). You must sow your work on te oter

More information

MATH745 Fall MATH745 Fall

MATH745 Fall MATH745 Fall MATH745 Fall 5 MATH745 Fall 5 INTRODUCTION WELCOME TO MATH 745 TOPICS IN NUMERICAL ANALYSIS Instructor: Dr Bartosz Protas Department of Matematics & Statistics Email: bprotas@mcmasterca Office HH 36, Ext

More information

IEOR 165 Lecture 10 Distribution Estimation

IEOR 165 Lecture 10 Distribution Estimation IEOR 165 Lecture 10 Distribution Estimation 1 Motivating Problem Consider a situation were we ave iid data x i from some unknown distribution. One problem of interest is estimating te distribution tat

More information

Polynomial Interpolation

Polynomial Interpolation Capter 4 Polynomial Interpolation In tis capter, we consider te important problem of approximatinga function fx, wose values at a set of distinct points x, x, x,, x n are known, by a polynomial P x suc

More information

New Distribution Theory for the Estimation of Structural Break Point in Mean

New Distribution Theory for the Estimation of Structural Break Point in Mean New Distribution Teory for te Estimation of Structural Break Point in Mean Liang Jiang Singapore Management University Xiaou Wang Te Cinese University of Hong Kong Jun Yu Singapore Management University

More information

How to combine M-estimators to estimate. quantiles and a score function.

How to combine M-estimators to estimate. quantiles and a score function. How to combine M-estimators to estimate quantiles and a score function. Andrzej S. Kozek Department of Statistics, C5C, Macquarie University Sydney, NSW 2109, Australia October 29, 2004 Abstract. In Kozek

More information

Boosting Kernel Density Estimates: a Bias Reduction. Technique?

Boosting Kernel Density Estimates: a Bias Reduction. Technique? Boosting Kernel Density Estimates: a Bias Reduction Tecnique? Marco Di Marzio Dipartimento di Metodi Quantitativi e Teoria Economica, Università di Cieti-Pescara, Viale Pindaro 42, 65127 Pescara, Italy

More information

LAPLACIAN MATRIX LEARNING FOR SMOOTH GRAPH SIGNAL REPRESENTATION

LAPLACIAN MATRIX LEARNING FOR SMOOTH GRAPH SIGNAL REPRESENTATION LAPLACIAN MATRIX LEARNING FOR SMOOTH GRAPH SIGNAL REPRESENTATION Xiaowen Dong, Dorina Tanou, Pascal Frossard and Pierre Vandergeynst Media Lab, MIT, USA xdong@mit.edu Signal Processing Laboratories, EPFL,

More information

Volume 29, Issue 3. Existence of competitive equilibrium in economies with multi-member households

Volume 29, Issue 3. Existence of competitive equilibrium in economies with multi-member households Volume 29, Issue 3 Existence of competitive equilibrium in economies wit multi-member ouseolds Noriisa Sato Graduate Scool of Economics, Waseda University Abstract Tis paper focuses on te existence of

More information

Homework 1 Due: Wednesday, September 28, 2016

Homework 1 Due: Wednesday, September 28, 2016 0-704 Information Processing and Learning Fall 06 Homework Due: Wednesday, September 8, 06 Notes: For positive integers k, [k] := {,..., k} denotes te set of te first k positive integers. Wen p and Y q

More information

Local Rank Inference for Varying Coefficient Models

Local Rank Inference for Varying Coefficient Models Local Rank Inference for Varying Coefficient Models Lan WANG, BoAI, and Runze LI By allowing te regression coefficients to cange wit certain covariates, te class of varying coefficient models offers a

More information

Estimating Peak Bone Mineral Density in Osteoporosis Diagnosis by Maximum Distribution

Estimating Peak Bone Mineral Density in Osteoporosis Diagnosis by Maximum Distribution International Journal of Clinical Medicine Researc 2016; 3(5): 76-80 ttp://www.aascit.org/journal/ijcmr ISSN: 2375-3838 Estimating Peak Bone Mineral Density in Osteoporosis Diagnosis by Maximum Distribution

More information

Adaptive Neural Filters with Fixed Weights

Adaptive Neural Filters with Fixed Weights Adaptive Neural Filters wit Fixed Weigts James T. Lo and Justin Nave Department of Matematics and Statistics University of Maryland Baltimore County Baltimore, MD 150, U.S.A. e-mail: jameslo@umbc.edu Abstract

More information

Order of Accuracy. ũ h u Ch p, (1)

Order of Accuracy. ũ h u Ch p, (1) Order of Accuracy 1 Terminology We consider a numerical approximation of an exact value u. Te approximation depends on a small parameter, wic can be for instance te grid size or time step in a numerical

More information

Learning based super-resolution land cover mapping

Learning based super-resolution land cover mapping earning based super-resolution land cover mapping Feng ing, Yiang Zang, Giles M. Foody IEEE Fellow, Xiaodong Xiuua Zang, Siming Fang, Wenbo Yun Du is work was supported in part by te National Basic Researc

More information

INFINITE ORDER CROSS-VALIDATED LOCAL POLYNOMIAL REGRESSION. 1. Introduction

INFINITE ORDER CROSS-VALIDATED LOCAL POLYNOMIAL REGRESSION. 1. Introduction INFINITE ORDER CROSS-VALIDATED LOCAL POLYNOMIAL REGRESSION PETER G. HALL AND JEFFREY S. RACINE Abstract. Many practical problems require nonparametric estimates of regression functions, and local polynomial

More information

(a) At what number x = a does f have a removable discontinuity? What value f(a) should be assigned to f at x = a in order to make f continuous at a?

(a) At what number x = a does f have a removable discontinuity? What value f(a) should be assigned to f at x = a in order to make f continuous at a? Solutions to Test 1 Fall 016 1pt 1. Te grap of a function f(x) is sown at rigt below. Part I. State te value of eac limit. If a limit is infinite, state weter it is or. If a limit does not exist (but is

More information

VARIANCE ESTIMATION FOR COMBINED RATIO ESTIMATOR

VARIANCE ESTIMATION FOR COMBINED RATIO ESTIMATOR Sankyā : Te Indian Journal of Statistics 1995, Volume 57, Series B, Pt. 1, pp. 85-92 VARIANCE ESTIMATION FOR COMBINED RATIO ESTIMATOR By SANJAY KUMAR SAXENA Central Soil and Water Conservation Researc

More information

Efficient algorithms for for clone items detection

Efficient algorithms for for clone items detection Efficient algoritms for for clone items detection Raoul Medina, Caroline Noyer, and Olivier Raynaud Raoul Medina, Caroline Noyer and Olivier Raynaud LIMOS - Université Blaise Pascal, Campus universitaire

More information

New families of estimators and test statistics in log-linear models

New families of estimators and test statistics in log-linear models Journal of Multivariate Analysis 99 008 1590 1609 www.elsevier.com/locate/jmva ew families of estimators and test statistics in log-linear models irian Martín a,, Leandro Pardo b a Department of Statistics

More information

ERROR BOUNDS FOR THE METHODS OF GLIMM, GODUNOV AND LEVEQUE BRADLEY J. LUCIER*

ERROR BOUNDS FOR THE METHODS OF GLIMM, GODUNOV AND LEVEQUE BRADLEY J. LUCIER* EO BOUNDS FO THE METHODS OF GLIMM, GODUNOV AND LEVEQUE BADLEY J. LUCIE* Abstract. Te expected error in L ) attimet for Glimm s sceme wen applied to a scalar conservation law is bounded by + 2 ) ) /2 T

More information

Math 1241 Calculus Test 1

Math 1241 Calculus Test 1 February 4, 2004 Name Te first nine problems count 6 points eac and te final seven count as marked. Tere are 120 points available on tis test. Multiple coice section. Circle te correct coice(s). You do

More information

arxiv: v1 [math.oc] 18 May 2018

arxiv: v1 [math.oc] 18 May 2018 Derivative-Free Optimization Algoritms based on Non-Commutative Maps * Jan Feiling,, Amelie Zeller, and Cristian Ebenbauer arxiv:805.0748v [mat.oc] 8 May 08 Institute for Systems Teory and Automatic Control,

More information

Error estimates for a semi-implicit fully discrete finite element scheme for the mean curvature flow of graphs

Error estimates for a semi-implicit fully discrete finite element scheme for the mean curvature flow of graphs Interfaces and Free Boundaries 2, 2000 34 359 Error estimates for a semi-implicit fully discrete finite element sceme for te mean curvature flow of graps KLAUS DECKELNICK Scool of Matematical Sciences,

More information

The derivative function

The derivative function Roberto s Notes on Differential Calculus Capter : Definition of derivative Section Te derivative function Wat you need to know already: f is at a point on its grap and ow to compute it. Wat te derivative

More information

MTH 119 Pre Calculus I Essex County College Division of Mathematics Sample Review Questions 1 Created April 17, 2007

MTH 119 Pre Calculus I Essex County College Division of Mathematics Sample Review Questions 1 Created April 17, 2007 MTH 9 Pre Calculus I Essex County College Division of Matematics Sample Review Questions Created April 7, 007 At Essex County College you sould be prepared to sow all work clearly and in order, ending

More information

Generic maximum nullity of a graph

Generic maximum nullity of a graph Generic maximum nullity of a grap Leslie Hogben Bryan Sader Marc 5, 2008 Abstract For a grap G of order n, te maximum nullity of G is defined to be te largest possible nullity over all real symmetric n

More information

arxiv: v3 [cs.ds] 4 Aug 2017

arxiv: v3 [cs.ds] 4 Aug 2017 Non-preemptive Sceduling in a Smart Grid Model and its Implications on Macine Minimization Fu-Hong Liu 1, Hsiang-Hsuan Liu 1,2, and Prudence W.H. Wong 2 1 Department of Computer Science, National Tsing

More information

Handling Missing Data on Asymmetric Distribution

Handling Missing Data on Asymmetric Distribution International Matematical Forum, Vol. 8, 03, no. 4, 53-65 Handling Missing Data on Asymmetric Distribution Amad M. H. Al-Kazale Department of Matematics, Faculty of Science Al-albayt University, Al-Mafraq-Jordan

More information

Chapter 5 FINITE DIFFERENCE METHOD (FDM)

Chapter 5 FINITE DIFFERENCE METHOD (FDM) MEE7 Computer Modeling Tecniques in Engineering Capter 5 FINITE DIFFERENCE METHOD (FDM) 5. Introduction to FDM Te finite difference tecniques are based upon approximations wic permit replacing differential

More information

Block Bootstrap Prediction Intervals for Autoregression

Block Bootstrap Prediction Intervals for Autoregression Department of Economics Working Paper Block Bootstrap Prediction Intervals for Autoregression Jing Li Miami University 2013 Working Paper # - 2013-02 Block Bootstrap Prediction Intervals for Autoregression

More information

POLYNOMIAL AND SPLINE ESTIMATORS OF THE DISTRIBUTION FUNCTION WITH PRESCRIBED ACCURACY

POLYNOMIAL AND SPLINE ESTIMATORS OF THE DISTRIBUTION FUNCTION WITH PRESCRIBED ACCURACY APPLICATIONES MATHEMATICAE 36, (29), pp. 2 Zbigniew Ciesielski (Sopot) Ryszard Zieliński (Warszawa) POLYNOMIAL AND SPLINE ESTIMATORS OF THE DISTRIBUTION FUNCTION WITH PRESCRIBED ACCURACY Abstract. Dvoretzky

More information

OSCILLATION OF SOLUTIONS TO NON-LINEAR DIFFERENCE EQUATIONS WITH SEVERAL ADVANCED ARGUMENTS. Sandra Pinelas and Julio G. Dix

OSCILLATION OF SOLUTIONS TO NON-LINEAR DIFFERENCE EQUATIONS WITH SEVERAL ADVANCED ARGUMENTS. Sandra Pinelas and Julio G. Dix Opuscula Mat. 37, no. 6 (2017), 887 898 ttp://dx.doi.org/10.7494/opmat.2017.37.6.887 Opuscula Matematica OSCILLATION OF SOLUTIONS TO NON-LINEAR DIFFERENCE EQUATIONS WITH SEVERAL ADVANCED ARGUMENTS Sandra

More information

Symmetry Labeling of Molecular Energies

Symmetry Labeling of Molecular Energies Capter 7. Symmetry Labeling of Molecular Energies Notes: Most of te material presented in tis capter is taken from Bunker and Jensen 1998, Cap. 6, and Bunker and Jensen 2005, Cap. 7. 7.1 Hamiltonian Symmetry

More information

Printed Name: Section #: Instructor:

Printed Name: Section #: Instructor: Printed Name: Section #: Instructor: Please do not ask questions during tis exam. If you consider a question to be ambiguous, state your assumptions in te margin and do te best you can to provide te correct

More information

Differential Calculus (The basics) Prepared by Mr. C. Hull

Differential Calculus (The basics) Prepared by Mr. C. Hull Differential Calculus Te basics) A : Limits In tis work on limits, we will deal only wit functions i.e. tose relationsips in wic an input variable ) defines a unique output variable y). Wen we work wit

More information

LEAST-SQUARES FINITE ELEMENT APPROXIMATIONS TO SOLUTIONS OF INTERFACE PROBLEMS

LEAST-SQUARES FINITE ELEMENT APPROXIMATIONS TO SOLUTIONS OF INTERFACE PROBLEMS SIAM J. NUMER. ANAL. c 998 Society for Industrial Applied Matematics Vol. 35, No., pp. 393 405, February 998 00 LEAST-SQUARES FINITE ELEMENT APPROXIMATIONS TO SOLUTIONS OF INTERFACE PROBLEMS YANZHAO CAO

More information

Click here to see an animation of the derivative

Click here to see an animation of the derivative Differentiation Massoud Malek Derivative Te concept of derivative is at te core of Calculus; It is a very powerful tool for understanding te beavior of matematical functions. It allows us to optimize functions,

More information

5 Ordinary Differential Equations: Finite Difference Methods for Boundary Problems

5 Ordinary Differential Equations: Finite Difference Methods for Boundary Problems 5 Ordinary Differential Equations: Finite Difference Metods for Boundary Problems Read sections 10.1, 10.2, 10.4 Review questions 10.1 10.4, 10.8 10.9, 10.13 5.1 Introduction In te previous capters we

More information

Digital Filter Structures

Digital Filter Structures Digital Filter Structures Te convolution sum description of an LTI discrete-time system can, in principle, be used to implement te system For an IIR finite-dimensional system tis approac is not practical

More information

Exercises for numerical differentiation. Øyvind Ryan

Exercises for numerical differentiation. Øyvind Ryan Exercises for numerical differentiation Øyvind Ryan February 25, 2013 1. Mark eac of te following statements as true or false. a. Wen we use te approximation f (a) (f (a +) f (a))/ on a computer, we can

More information

An Empirical Bayesian interpretation and generalization of NL-means

An Empirical Bayesian interpretation and generalization of NL-means Computer Science Tecnical Report TR2010-934, October 2010 Courant Institute of Matematical Sciences, New York University ttp://cs.nyu.edu/web/researc/tecreports/reports.tml An Empirical Bayesian interpretation

More information

Department of Statistics & Operations Research, Aligarh Muslim University, Aligarh, India

Department of Statistics & Operations Research, Aligarh Muslim University, Aligarh, India Open Journal of Optimization, 04, 3, 68-78 Publised Online December 04 in SciRes. ttp://www.scirp.org/ournal/oop ttp://dx.doi.org/0.436/oop.04.34007 Compromise Allocation for Combined Ratio Estimates of

More information

Te comparison of dierent models M i is based on teir relative probabilities, wic can be expressed, again using Bayes' teorem, in terms of prior probab

Te comparison of dierent models M i is based on teir relative probabilities, wic can be expressed, again using Bayes' teorem, in terms of prior probab To appear in: Advances in Neural Information Processing Systems 9, eds. M. C. Mozer, M. I. Jordan and T. Petsce. MIT Press, 997 Bayesian Model Comparison by Monte Carlo Caining David Barber D.Barber@aston.ac.uk

More information

Exam 1 Review Solutions

Exam 1 Review Solutions Exam Review Solutions Please also review te old quizzes, and be sure tat you understand te omework problems. General notes: () Always give an algebraic reason for your answer (graps are not sufficient),

More information

INTRODUCTION AND MATHEMATICAL CONCEPTS

INTRODUCTION AND MATHEMATICAL CONCEPTS INTODUCTION ND MTHEMTICL CONCEPTS PEVIEW Tis capter introduces you to te basic matematical tools for doing pysics. You will study units and converting between units, te trigonometric relationsips of sine,

More information

Copyright c 2008 Kevin Long

Copyright c 2008 Kevin Long Lecture 4 Numerical solution of initial value problems Te metods you ve learned so far ave obtained closed-form solutions to initial value problems. A closedform solution is an explicit algebriac formula

More information

A Simple Matching Method for Estimating Sample Selection Models Using Experimental Data

A Simple Matching Method for Estimating Sample Selection Models Using Experimental Data ANNALS OF ECONOMICS AND FINANCE 6, 155 167 (2005) A Simple Matcing Metod for Estimating Sample Selection Models Using Experimental Data Songnian Cen Te Hong Kong University of Science and Tecnology and

More information

Material for Difference Quotient

Material for Difference Quotient Material for Difference Quotient Prepared by Stepanie Quintal, graduate student and Marvin Stick, professor Dept. of Matematical Sciences, UMass Lowell Summer 05 Preface Te following difference quotient

More information

Kernel estimates of nonparametric functional autoregression models and their bootstrap approximation

Kernel estimates of nonparametric functional autoregression models and their bootstrap approximation Electronic Journal of Statistics Vol. (217) ISSN: 1935-7524 Kernel estimates of nonparametric functional autoregression models and teir bootstrap approximation Tingyi Zu and Dimitris N. Politis Department

More information

Consider a function f we ll specify which assumptions we need to make about it in a minute. Let us reformulate the integral. 1 f(x) dx.

Consider a function f we ll specify which assumptions we need to make about it in a minute. Let us reformulate the integral. 1 f(x) dx. Capter 2 Integrals as sums and derivatives as differences We now switc to te simplest metods for integrating or differentiating a function from its function samples. A careful study of Taylor expansions

More information

Function Composition and Chain Rules

Function Composition and Chain Rules Function Composition and s James K. Peterson Department of Biological Sciences and Department of Matematical Sciences Clemson University Marc 8, 2017 Outline 1 Function Composition and Continuity 2 Function

More information

GRID CONVERGENCE ERROR ANALYSIS FOR MIXED-ORDER NUMERICAL SCHEMES

GRID CONVERGENCE ERROR ANALYSIS FOR MIXED-ORDER NUMERICAL SCHEMES GRID CONVERGENCE ERROR ANALYSIS FOR MIXED-ORDER NUMERICAL SCHEMES Cristoper J. Roy Sandia National Laboratories* P. O. Box 5800, MS 085 Albuquerque, NM 8785-085 AIAA Paper 00-606 Abstract New developments

More information

THE hidden Markov model (HMM)-based parametric

THE hidden Markov model (HMM)-based parametric JOURNAL OF L A TEX CLASS FILES, VOL. 6, NO. 1, JANUARY 2007 1 Modeling Spectral Envelopes Using Restricted Boltzmann Macines and Deep Belief Networks for Statistical Parametric Speec Syntesis Zen-Hua Ling,

More information

DEPARTMENT MATHEMATIK SCHWERPUNKT MATHEMATISCHE STATISTIK UND STOCHASTISCHE PROZESSE

DEPARTMENT MATHEMATIK SCHWERPUNKT MATHEMATISCHE STATISTIK UND STOCHASTISCHE PROZESSE U N I V E R S I T Ä T H A M B U R G A note on residual-based empirical likeliood kernel density estimation Birte Musal and Natalie Neumeyer Preprint No. 2010-05 May 2010 DEPARTMENT MATHEMATIK SCHWERPUNKT

More information

MVT and Rolle s Theorem

MVT and Rolle s Theorem AP Calculus CHAPTER 4 WORKSHEET APPLICATIONS OF DIFFERENTIATION MVT and Rolle s Teorem Name Seat # Date UNLESS INDICATED, DO NOT USE YOUR CALCULATOR FOR ANY OF THESE QUESTIONS In problems 1 and, state

More information

3.1 Extreme Values of a Function

3.1 Extreme Values of a Function .1 Etreme Values of a Function Section.1 Notes Page 1 One application of te derivative is finding minimum and maimum values off a grap. In precalculus we were only able to do tis wit quadratics by find

More information

ADCP MEASUREMENTS OF VERTICAL FLOW STRUCTURE AND COEFFICIENTS OF FLOAT IN FLOOD FLOWS

ADCP MEASUREMENTS OF VERTICAL FLOW STRUCTURE AND COEFFICIENTS OF FLOAT IN FLOOD FLOWS ADCP MEASUREMENTS OF VERTICAL FLOW STRUCTURE AND COEFFICIENTS OF FLOAT IN FLOOD FLOWS Yasuo NIHEI (1) and Takeiro SAKAI (2) (1) Department of Civil Engineering, Tokyo University of Science, 2641 Yamazaki,

More information