الفهرس | Only 14 pages are availabe for public view |
Abstract The staggering proliferation of machine learning models for regression is recognized as a convenient way of improving the estimation of empirical models. Among most popular machine learning-based methods that are used to build the prediction model is the support vector regression (SVR). However, the combination of principle component analysis (PCA) and factor analysis (FA) as a feature dimension reduction method with support vector regression (SVR) is also possible. We investigate the competence of support vector regression (SVR) after employing principle component analysis (PCA) and factor analysis (FA) to explore the possibility of reducing data and yielding more accurate estimation.The objective of the study is to compare f-support vector regression and v-support vector regression models with applyingprincipal component analysis and factor analysisfor simulation study and applied the results on real data from Iraq. The performance criteria for comparison are root mean squared error (RMSE) and squared correlation coefficient (R2). A simulation study and Renal Failure (RF) data of support vector regression (SVR) optimized by kernel functions (linear, polynomial, radial, sigmoid) has been done using R programming to compare the behavior of f-SVR and v-SVR models on different sample sizes.The structure of this dissertation is organized into five chapters as follows: Chapter Ipresents an introductory to the dissertation.Chapter II reviews the previous work related to principal component analysis, factor analysis, and support vector regression.Chapter III includes some definitions that will be considered in the next chapters of this thesis. It also presents an overview of data reduction, principal component analysis, factor analysis, and support vector regression. Chapter IV presents the simulation study using R software program 3.2.5 in which principal component analysis and factor analysis are employed for data reduction under two types of support vector regression models for different kernels function using different sample sizes.Chapter V gathers the real medical data set collected from Hiwahospital regarding renal failure as an application. It also includes important sections on the results of using the methods used in the previous chapters in order to be discussed and draws a conclusion of the study |