SCHO OL OF STATISTIC S UNIVERSITY OF THE PHILIPPINES DILIMAN WORKING PAPER SERIES SEMIPARAMETRIC PRINCIPAL COMPONENTS POISSON REGRESSION ON CLUSTERED DATA by Erniel B. Barrios and Kristina Celene M. Manalaysay UPSS Working Paper No. 2011-06 November 2011 School of Statistics Ramon Magsaysay Avenue U.P. Diliman, Quezon City Telefax: 928-08-81 Email: updstat@yahoo.com SEMIPARAMETRIC PRINCIPAL COMPONENTS POISSON REGRESSION ON CLUSTERED DATA Erniel B. Barrios School of Statistics, University of the Philippines Diliman Kristina Celene M. Manalaysay School of Statistics, University of the Philippines Diliman ABSTRACT Clustering of observations and interdependency of predictors are two common problems in modeling count data with multivariate predictors. We propose to use principal components of predictors to mitigate the multicollinearity problem. To abate information losses due to dimension reduction, a semiparametric link between the count dependent variable and the principal components is postulated. Clustering of observations is accounted into the model as a random component and the model is estimated via the backfitting algorithm. A simulation study illustrates the advantages of the proposed model over standard poisson regression in a wide range of simulation scenarios. Keywords: semiparametricpoissonregression, clustered data, multicollinearity, principal components analysis 1. Introduction In many diverse fields, outcomes of certain phenomena are measured using indicators that possess the characteristics of poisson events, e.g., prevalence of a disease, number of customers patronizing products/services, number of student enrollees. Poisson regression is used to characterizesuch dataand in predicting the average number of instances an event occurs, conditional on one or more factors. Ruru and Barrios (2003) demonstrated using malaria data that poisson regression is advantageous over classical regression in modelling count data. Classical regression analysis requires more predictors to achieve as much predictive ability as poisson regression. Spatial aggregation causes certain poisson events to manifest clustering.The spread of AH1N1 is influenced by determinants leadings towards vulnerability of individuals in the same community, this may be different from those causing vulnerability of other individuals from a different community. Clusters may still be independent but members of the same cluster (or neighborhood) are necessarily dependent since there is some spatial endowment commonly shared among units that formed the cluster. Classical statistical inference assumes independence of observations, i.e., data are independently collected on similar, homogenous units. This assumption is not necessarily true for clustered data. Thus, in analyzing clustered data with methods that implicitly consider independence of observations may yield incorrect analyseson the dynamics of the events/phenomena being characterized. Predictors that explain occurrence of poisson events within the cluster can also be naturally correlated. The interdependence among predictors usually causes problems in statistical inference involving linear models. Themulticollinearityproblem exists when two or more explanatory variables in a regression model are highly correlatedimplicating the inefficiency of ordinary least squares estimates of the regression coefficients. As an illustration, consider income and educational attainment as predictors of political preference. Income and educational attainment are structurally correlated since income varies according to the level of educational attainment of an individual. The presence of multicollinearity in a statistical model inflates the standard error of the estimated coefficients, resulting in unreliablecharacterization of the coefficients (Curto and Pinto, 2007). It further weakens in sensitivity of the dependent variable on changes in independent variables and makes it difficult to assess the relative importance of the independent variables in the model. There are several solutions to the multicollinearity problem. For example, instead of individual predictors, some important principal componentsare used in the model. In the presence of multicollinearity, the design matrix becomes ill-conditioned if not singular and hence, principal components analysis transforms correlated variables into fewer independent components. Since Principal Components Regression (PCR) uses only a subset of the principal components, there is a loss of information resulting to thedeterioration of the predictive ability of the estimated regression function compared to the model that usesall the individual predictors (Dunteman, 1989). It is also possible that the use of a subset of principal components can result to bias in the assessment of the relative importance of a predictor in explaining the dependent variable. The lost information in principal components regression can be recovered by allowing flexibility on the functional relationship between the dependent variable and the principal components, (Barrios and Umali, 2011). In nonparametric regression, the functional form of the link between the dependent and independent variables is allowed to be flexible with only the requirement of smoothness of the function incorporated into the objective function of the estimation. With a flexible functional form, the principal components can have a more accurate characterization of the variation of the dependent variable, hence improving its predictive ability. We postulate an additive combination of nonparametric functions on principal componentsand random effects in a regression model with measurements of poisson events as the dependent variable. This semiparametricpoisson regression model can be used in characterizing high dimensional clustered data. Clustering effect is accounted into the model through a random intercept term. Dimension reduction is achieved through principal components and due to the inherent deterioration in model fit due to dimension reduction, the covariate effect summarized in terms of the principal components will be postulated as nonparametric functions. 2. Some Modeling Strategies Classical linear regression assumes continuous dependent variable and will lead to inefficient, inconsistent and biased estimates when used in count dependent variable. Poisson regression is appropriate in modeling with count dependent variable data.Even if poisson regression can be approximated by classical linear regression, e.g., large sample size, poisson regression is advantageous over classical linear regression since it usually requires fewer predictorsto achieve a good fit, as demonstrated in the study of malaria incidence by Ruru and Barrios (2003). Nelder and Wedderburn(1972) introduced the generalized linear models (GLM) to relax some of the classical assumptions of a linear model. The model is given by where exponential family and systematic component for every i, Yibelongs to the is a function that links the random component Yi to the . These are developed for regression models with non-normal dependent variables; special cases include poisson regression where Y is a count variable and logistic regression where Y is a binary outcome. Demidenko (2007) compared the following models for clustered data: (1) ordinary poisson regression, which ignores intracluster correlation, (2) poisson regression with fixed cluster-specific intercepts, (3) a generalized estimating equations approach with an equicorrelation matrix, (4) an exact generalized estimating equations approach with an exact covariance matrix, and (5) maximum likelihood. All five methods lead to consistent estimates of slopes but have yield varying efficiency levels speciallyfor unbalanced data. Poisson regression assumes heterogeneous mean that is expressed as a linear combination of explanatory variables. Since the parameter λ is positive, it is convenient to express this parameter through an exponential function, , where β is an mx1 vector of regression parameters and xi is an mx1 vector of explanatory variables or covariates, i = 1,2,…,n. In the context of generalized linear models, poisson regression has log link because , (Demidenko, 2004).Poisson model is useful for clustered data since the cluster-specific intercepts may be eliminated and can be viewed as a limiting maximum likelihood estimates when the variance of the intercepts approaches infinity (Demidenko, 2007). Principal components analysis (PCA) transforms a set of p correlated variables into uncorrelated linear combinations called principal components (PC’s). PCA rotates the original variable space to a point where the variance of the new variate is maximized. Since the PC’s are ranked by order of explained variance, the last PC’s have the smallest variance but it is through the last PC’s that the relationship of the independent variables to the dependent variable are determined, that is, the variables with high loadings on the last PC’s are proven to be highly correlated (Jolliffe, 2002). In modeling where only a subset of the PC’s areused, there is substantial loss of information. Typically, only a subset of the PC’s is included in regression modeling, though there is no universally acceptable procedure yet to determine the PC’s to retain (Draper and Smith, 1981). Filzmoser and Croux (2002) proposed a procedure that simultaneously chooses the components while model fit is optimized. Montgomery and Peck (1982) and Dunteman (1989) justified why the principal components with low eigenvalues are not included in the model. Since the variance of the estimator (of ) is a linear combination of the reciprocal of the eigenvalues, inclusion of one or more components with small eigenvalues in the model yield high variance of . Nevertheless, Marx and Smith (1990)noted that given specific theoretical models oriented towards parameter estimation, principal component regression can yield desirable (maximum) variance property with minimal bias. The bias and the lost information in principal component regression should be addressed. For instance, nonparametric smoothing techniques which aim to provide a strategy in modeling the relationships between variables without specifying any particular form for the underlying regression function may be considered.When several covariates are present, Hastie and Tibshirani (1986) proposed to extend the idea of linear regression into a flexible form known as generalized additive model (GAM). The regression model is given by where are nonparametric components. Additive models assume nonparametric smoothing splines for each predictor in regression models. Hastie and Tibshiranie (1990) suggested that additive models are used as initial procedure to locate the patterns and behavior of the predictors relative to the response, suggesting a possible parametric form for which to model Y at a later stage. Alvarez and Pardinas (2009) formulated a censored regression model with additive effects of the covariates. The additive model sped up computation in the inference process and yieldmore promising results over a class of linear models especially where there is violation in the linearity assumption. 3. Methodology Clustered data is characterized by homogeneity of elements within the cluster and heterogeneity between clusters. This case would require a model structure that accounts for variation across the clusters, while accounting for similarity of subjects within the cluster.The choice of covariates or predictors that can sufficiently account for between-cluster variability and within-cluster homogeneity is often complicated resulting to very long list. This would necessarily lead to a high dimensional data type and often, this can invite other confounding problems. High dimensional data becomes prone to multicollinearity and suffers from the curse of dimensionality. Multicollinearitycan be a crucial issue in modeling. However, there are various approaches to mitigate its ill-effects, e.g., transform the individual predictors into linear combinations, the combinations are chosen so that they are independent, yet it contains the maximum amount of variance that the original predictorscontain. With linear combinations as predictors instead of the individual variables in a linear model, the predictive ability could suffer. On the other hand, high dimensionality of predictors, multicollinearity, discreteness of the dependent variable (count data), and the clustering usually associated with poisson process that generates the count data can posed many complications in many model structures. Thus, we offer a solution that can potentially alleviate the predictive ability of the model through a nonparametric postulated link function in poisson regression for count data. 3.1 Postulated Model De Vera (2010) proposed asemiparametricpoisson regression model for spatially clustered count data given npredetermined clusters with n k observationsin eachcluster given by (1) Equivalently, Model 1 can also be written as: (2) where are the explanatory variables are smoothfunctions of are clusterrandom intercepts isthe error term isthe response variable As the number of predictorsincreases, the likelihood of themulticollinearity problem also increases and either can cause potential problems in estimating Model 2. We propose to use principal components of the predictors instead of the individual variables in the model. This however leads to bias since some information on the predictors can be lost in the process. Thus, the link function is postulated as a nonparametric function resulting to the following semiparametric model: (3) where are cluster random intercepts is the score for jth principal component on the ith observation is the error term is the response variable The flexibility in form of the nonparametric function of the principal components will compensate for the bias in the estimation of the regression coefficients due to the information lost by selecting only the most important principal components to be included as predictors in the model. The random intercepts will account for cluster differences (homogeneous within a cluster) and the possible peculiarities in the model caused by clustering. 3.2 Estimation Procedures The principle of backfitting is used to estimate the parametric and nonparametric components of the model.Assuming additivity of Model (3), the parametric and nonparametric frameworks are imbedded into the backfittingg algorithm to estimate the parameters and the nonparametric components of the model. Two approaches are presented in this section.In Method 1, the backfitting algorithm estimates the nonparametric part first, and then the parametric part is estimated from the residuals. In Method 2, ordinary poisson regression is usedbut with principal components as predictors. 3.2.1 Method 1: Backfitting of the Semiparametric Model After the extraction of principal components, the parametric and nonparametric parts of the model are estimated iteratively in the context of backfitting. The nonparametric functions of the principal components are estimated through spline smoothing. Smoothing splines are used to fit the nonparametric part of the model. Consider a simple additive regression model Y = f(X) + ε, where E(ε) = 0 and Var(ε) = σ2. We want to estimate the function fas a solution to the penalized least squares problem given by given a value of the smoothing parameter >0. The first term measures the goodness of fit and the second term served as the penalty for lack of smoothness in f due to the interpolation in the first term.The smoothing parameter controls the tradeoff between smoothness and goodness of fit. Largevalues of emphasizes smoothness of f over model fit, while small values put higher leverage on model fit rather than on smoothness of f. As , the solution to f is an interpolation of the data points.The choice of the value of smoothing parameter is optimized through the generalized cross validation (GCV). is chosen to minimize the generalized cross validation mean squared error given by , see (Golub, et. al., 1979), (Hardle, et. al., 2004) for further details. The partial residual e i is computed and used to estimate the random intercepts (with a priori information on clustering of observations) through methods like maximum likelihood methods, EM algorithm, see for example (Seidel, 2011). Spline smoothing and mixed model estimation in the backfitting framework are then iterated until convergence, see (Opsomer, 2000) for some optimal properties of the backfitting estimators. 3.2.2 Method 2: Ordinary Poisson Regression on the Principal Components With poisson link function, a general linear model (GLM) with the principal components as predictors was also estimated. In GLM, the outcome Y is assumed to be generated from a particular distribution in the exponential family, in this case, from a poisson distribution. The heterogeneous mean, μ, of the distribution depends on the independent variables, in this case, the PC’s, through: (4) These two methods are compared with Ordinary Poisson Regression on the original predictors in terms of their predictive ability. 3.3 Simulation Studies We conduct a simulation study that covers the features of typical data to be analyzed using Model (3), i.e., either it includes high dimensional predictors, or that the multicollinearity problem is present, or both. We then compared the predictive ability of Methods 1 and 2 with ordinary poisson regression with the individual variables as predictors. The simulation scenarios are summarized in Table 1: Table 1 Summary of Simulation Scenarios Number of Variables/ Expected Number of PC’s Few (5) Many (30) - Single Pattern Many (30) - Three Patterns Multicollinearity Absence Strong Sample Size 50 100 Number of Clusters 5 10 Model Fit Good Poor Starting with five X’s, the rest were simulated as an additive combination of a function of another X and the random error term. The initial set of predictors are simulated from: (5) The error term is distributed asN(0,1), the multiplier of the error term inducesmulticollinearityamong the predictors, i.e., higher multiplier implies absence of multicollinearitywhile lower multiplier indicates presence of multicollinearity.The twenty five other predictors were generated to fulfil various multicollinearity structures. The response variable was computed as the linear combination of the X’s and added with a cluster mean (from the normal distribution) and an error term. The means of the normal distribution from where the cluster means were simulated were spread thoroughly to differentiate the clusters. A multiplier to the error term is included to altermodel fit (large multiplier implies poor model fit, small multiplier implies good model fit). A model usually fits the data well when the functional form used is correct and that thecorrect predictors are aptly accounted into the model. When the functional form of the model is incorrect or that there are missed out predictors, variation of the error term will dominate. Hence, to simulate misspecification, we magnified the error by multiplying it with a constant. Scenarios for varying sample sizes (50, 100) as well as varyingnumber of clusters (5,10) were generated to assess robustness of the proposed model. Furthermore, correlations among predictors that can result to absence of multicollinearity or severe multicollinearity were included. The number of variables can be 5 or 30, with the 30-variable scenarios further divided into single and three-patternmulticollinearitystucture. Single-pattern scenarios represent data with the variables being correlated with one variable, i.e., they are computed as a function of a particular variable only. Three-pattern scenarios, on the other hand, have a number of these “base” variables, from whom the rest of the variables are computed. These variations in pattern were incorporated to simulate varying predicted number of principal components. Each scenario is replicated 100 times. 4. Results and Discussion The predictive ability of the proposed model is assessed by comparing the mean absolute prediction error (MAPE)using estimation methods 1 and 2 as well as the MAPE obtained using Ordinary Poisson Regression (OPR) based on the original predictors, 4.1 Effect of Misspecification Error In empirical modeling, misspecification is very common especially when variables were measured in an ad hoc manner, i.e., not based on some theoretical foundation. Misspecification error is introduced by magnifying the error term, i.e., multiplying with a constant to inflate the variance. As the magnitude of the constant increases (equivalently, model fit worsens or that the extent of misspecification error increases), the mean absolute prediction error (MAPE) also increases. In the simulation study, “good” model fit is represented by linear equations where coefficient of determination is at least 60%, while “poor” model fit is associated with coefficient of determination lower than 60%. Simulation shows that whether misspecification error is present or not, the proposed semiparametric model always outperformed the other models in terms of predictive ability. In the absence of misspecification error, the proposed model is advantageous to ordinary poisson regression by about 10% in MAPE. Shifting from ordinary poisson regression to the parametric principal components regression, on the other hand, results in 65% increase in MAPE. This illustrates the effect of lost information due to the summarization of the predictors into principal components instead of using the individual predictors. The advantages of the proposed method are also observed even with the presence of misspecification errors. MAPE in the semiparametric model improved 9% over ordinary poisson regression. Furthermore, there is advantage of 48% in terms of MAPE in using ordinary poisson regression relative to the parametric model principal component model. The MAPE of the three models by nature of model fit are summarized in Table 2. Table 2. Comparison of MAPE for Varying Model Fit MAPE (%) Model Fit Semiparametric Model Parametric Model Good Poor 24.26 141.54 44.28 230.64 OPR on Original Predictors 26.88 156.23 Since substantial number of cases with “poor” model fit was included, subsequent results yield higher MAPE levels since predictive ability of those models with “good” fit were contaminated. The discussions then will focus on the comparison of the three models, rather than on the magnitude of MAPE. 4.2 Effect ofMulticollinearity In the absence of multicollinearity, the proposed model yields better prediction over ordinary poisson regression on the original predictors (OPR) where the MAPE is lower by 5%. The parametric principal component regression on the other hand, yields48% higher MAPE than OPR. This is explained as the effect of lost information due to selection only of the more important principal components. With multicollinearity present, similar ranges of MAPE can be observed from the proposed model with improvementby 14% over ordinary poisson regression. Again, the parametric model yield 52% higher MAPE relative to ordinary poisson regression on original variables.Principal Components Analysis is a procedure aimed at addressing multicollinearity, thus giving the proposed model advantage over ordinary poisson regression with original correlated variables as predictors. Thesemiparametric nature of the proposed model makes a more flexible regression and endowed advantages over the parametric model.The proposed model performs best among the three methods whether multicollinearityis present or not. The MAPE values are summarized in Table 3. Table 3. Comparison of MAPE with or withoutMulticollinearity MAPE (%) Multicollinearity Semiparametric Model Parametric Model OPR on Original Predictors Absence Presence 108.84 65.61 170.28 115.58 114.79 76.06 4.3 Effect of Sample Size We simulated two sample size values: 50 observations (small) and 100 observations (large).In small sample size,while ordinary poisson regressionyields better predictive ability among the three models, the proposed model is still within a comparable range. For large sample size, however, the proposed model is most advantageous, yielding 19% lower MAPE than ordinary poisson regression. This is consistent with the observation ofOsborne and Costello (2004) that in small samples, errors can easily occur particularly among multivariate techniques such as principal components analysis (eg extraction of erroneous principal components). For small samples, estimates of correlation among the predictors are relatively unstable, hence, the estimates of component loadings may not be accurate resulting to losses in information. This lost information is somehow retrieved by relaxing the parametric structure of the model and employ the more flexible nature of nonparametric regression. In Table 4, the parametric model is the most robust to sample size, with its MAPE changing only by 3% as sample size changes from 50 to 100. MAPE of the proposed model increased 23% as sample size increased, whereas MAPE of ordinary poisson regression increased 63%. In the parametric model, a sample of size 50 is already fairly large, hence no significant changes in MAPE is observed as the sample size is increased further to 100. While MAPE for the proposed model increases with the sample size, it still performed better than ordinary poisson regression and the parametric model. Table 4. Comparison of MAPE for Varying Sample Size MAPE (%) Sample Size Semiparametric Model Parametric Model 50 100 74.19 91.61 135.26 139.66 OPR on Original Predictors 69.59 113.51 4.4 Effect of Number of Clusters A priori information on clustering of observations can inform the modeler of possible sources of variation of the response variable. More clusters mean that there aremore basis for the attribution of heterogeneity of the observations. Furthermore, the proposed model included a random intercept term that basically explains the cluster differences. The simulation study illustrates the influence of the number of clusters in the proposed model. Increasing the number of clusters yield improvement on the predictive ability of the proposed model. In fewer clusters, MAPE of the semiparametric model improved by 7% compared toordinary poisson regression. On the other hand, MAPE increased by 49% in the parametric model from ordinary poisson regression. Similar trend can be observed for cases with more clusters, but the greater advantage of the semiparametric model over ordinary poisson regression (about 12% decline in MAPE) is observed in these cases. The above result is consistent with Arceneaux and Nickerson (2009) who recommended adding more clusters, not the individual observations, for increased efficiency in the analysis of clustered data. Table 5 provides details related to the effect of number of clusters on predictive ability of models in clustered data. Table 6 illustrates the implication of the number of clusters and number of observations per cluster in the analysis of clustered data. Table 5 Comparison of MAPE for Varying Cluster Count MAPE (%) Number of Clusters 5 10 Semiparametric Model Parametric Model 87.49 78.31 140.57 134.36 OPR on Original Predictors 94.19 88.91 Table 6. Comparison of MAPE for Varying Cluster Count and Cluster Size MAPE (5) Number of Clusters (Cluster Size) 5 (10) 5 (20) 10 (5) 10 (10) Semiparametric Model 79.40 95.58 68.98 87.64 Parametric Model 131.69 149.45 138.84 129.87 OPR on Original Predictors 64.20 124.18 74.98 102.84 4.5 Effect of Number of Variables While multicollinearity may appear even with few predictors, the chances of observing it in high dimensional data are higher. Scenarios with 5 (few) and 30 (many) variables were simulated in this study.In high dimensional data, principal component analysis is knownas an effective data reduction technique. Whether there are few or many variables, the semiparametric model still exhibitsbetter predictive ability.MAPE values are even lower in cases where there are 30variables. With fewer variables, the proposed model yields 17% lower MAPE compared toordinary poisson regression, whereas the parametric model yields 8% higher MAPE than that of ordinary poisson regression. Table 7indicates similar results for scenarios involving many variables. Table 7. Comparison of MAPE for Varying Number of Variables MAPE (%) Number of Variables 5 30 Semiparametric Model 81.53 78.19 Parametric Model 106.84 146.52 OPR on Original Predictors 98.36 83.37 For scenarios with thirty variables, presence of multicollinearityis further examined for the effect of only a single pattern of interdependence among the predictors and cases where there are three patterns of such interdependencies. The purpose of heterogeneity in interdependencies is to account for the number of principal components that need to be included in the model. The simulation study illustrates that predictive ability of the semiparametric model improves as the interdependencies become more complicated, i.e., more patterns of interdependencies. Compared to the ordinary poisson regression and the parametric model, the proposed model is still advantageous in terms of predictive ability. Table 8. Comparison of MAPE for Single- and Three-Pattern Scenarios MAPE (%) Number of Correlations Pattern 30 Single-Pattern 30 Three-Pattern Semiparametric Model 95.05 61.33 Parametric Model 180.60 112.43 OPR on Original Predictors 94.31 72.43 5.Conclusions The semiparametric principal component poisson regression model is aimed to characterize clustered data with high dimensional or correlated predictors.The proposed model provides solution to the modelling issues associated with high dimensional or correlated predictors while it mitigates the bias caused by lost information in dimension reduction. The predictive ability of the semiparametric model is superior in the presence of multicollinearity. While it is superior to parametric poisson principal component regression and ordinary poisson regression in any sample size, it is best in cases of large sample sizes. Although ordinary poissonregression performed better than the proposedmodel for small sample size, the performance of the proposed model is still at par with the other methods, and it is the more robust than ordinary poisson regression to changes in sample size. The proposed model and the corresponding estimation procedure is capable of mitigating the problem of multicollinearity by regressing on the principal components instead of on the original predictors. Furthermore, the nonparametric specification of the effect of principal components abate the potential reduction in the predictive ability of the model that is usually observed in principal components regression caused by loss in information from dimension reduction. REFERENCES Alvarez, J. and Pardinas, J., 2009, Additive Models in Censored Regression, Computational Statistics and Data Analysis, 53: 3490-3501. Arceneaux, K. and Nickerson, D., 2009, Modeling Certainty with Clustered Data: A Comparison of Methods, Political Analysis, 17:177-190. Barrios, E., and Umali, J., 2011, Nonparametric Principal Components Regression, Proceedings of the 58th World Congress of the International Statistical Institute. Curto, J. and Pinto, J., 2007, New Multicollinearity Indicators in Linear Regression Models, International Statistical Review, 75(1):114-121. De Vera, E., 2010, Semiparametric Poisson Regression for Clustered Data, Unpublished Thesis, School of Statistics, University of the Philippines Diliman. Draper, N. and Smith, H., 1981,Applied Regression Analysis, Second Edition. New York: John Wiley. Demidenko, E., 2007, Poisson Regression for Clustered Data, International Statistical Review, 75:96-113. Demidenko, E., 2004,Mixed Models Theory and Applications, New Jersey: John Wiley. Dunteman, J., 1989,Principal Component Analysis, Sage University Papers Series on Quantitative Applications in the Social Sciences, 07-069. Thousand Oaks, CA: Sage. Filzmoser, P. and Croax, C., 2002, Dimension Reduction of the Explanatory Variables in Multiple Linear Regression, Pliska. Stud. Math. Bulgaria, 29:1-12. Golub, G., Heath, M., and Wahba, G., 1979, Generalized Cross-Validation as a Method for Choosing a Good Ridge Parameter, Technometrics, 21(2): 215-223. Hastie, T. and Tibshirani R., 1990,Generalized Additive Models, London: Chapman and Hall. Hardle, W., Muller, M., Sperlich S., Werwatz, A., 2004,Nonparametric and Semiparametric Models, Berlin: Springer. Jolliffe, I.T., 2002,Principal Components Analysis, Special Edition. New York: Springer. Marx, B.D. and Smith, E.P., 1990, Principal Components Estimation for Generalized Linear Regression,Biometrika, 77(1):23-31. Montgomery, D.C. and Peck, E.A., 1982,Introduction to Linear Regression, New York: John Wiley. Nelder, J. and Wedderburn, R., 1972,Generalized Linear Models,Journal of the Royal Statistical Society, 135:370-384. Opsomer, J., 2000, Asymptotic Properties of Backfitting Estimators, Journal of Multivariate Analysis, 73:166-179. Osborne, J. and Costello, A., 2004, Sample Size and Subject to Item Ratio in Principal Components Analysis. Practical Assessment, Research & Evaluation, 9, Paper No. 11. Ruru, Y. and Barrios, E., 2003, Poisson Regression Models of Malaria Incidence in Jayapura, Indonesia, The Philippine Statistician, 52:27-38. Seidel, W., 2011, Mixture Models, in Lovric, M., ed., International Encyclopedia of Statistical Science, Springer Reference, pp. 827-829.