Dissertation > Excellent graduate degree dissertation topics show

Research on Kernel Function and Parameter Selection in Support Vector Machine and Its Application

Author: ZhuChunLei
Tutor: LiuYingAn
School: Nanjing Forestry University
Course: Applied Computer Technology
Keywords: Support Vector Machine Mixtures of Kernels Adaptive Hybrid Genetic Algorithm Parameter selection Outlier Detection
CLC: TP18
Type: Master's thesis
Year: 2011
Downloads: 177
Quote: 1
Read: Download Dissertation

Abstract


Support vector machines (Support Vector Machine, SVM) is a statistical learning theory developed in the 1990s, it is a classification and regression problems proposed by AT T Bell Laboratories V.Vapnik et al new machine learning method, which by means of optimization methods to solve machine learning problems, integrated optimal hyperplane Mercer kernel function, convex quadratic programming, sparse solution and slack variable number of technical, has a global optimum, simple structure ability to promote the advantages of very good results in pattern classification, regression analysis and probability density estimation. However, support vector machine is still in continuous development and improvement. In this paper, the SVM model, the structure of the kernel function, SVM parameter selection, and outlier detection in four areas studied. The details are as follows: First, an overview of the content of this study - statistical learning theory and support vector machine, to describe and compare the present study used more often several training algorithms and deformation algorithm for the contents of this article foreshadowing. Second, the introduction of fuzzy logic thinking, Gaussian kernel and fuzzy Sigmoid kernel function Gaussian kernel-based, and Sigmoid nuclear function, two new hybrid kernel function. Both mixed kernel function gathering local kernel function and global nuclear function, improve the learning accuracy of the SVM algorithm and reduce learning time. Experimental results show that the two mixed kernel function SVM classification accuracy or classification time is better than the traditional SVM algorithm based on a single kernel function. Third, the basis of the traditional genetic algorithm, gradient algorithm, proposed an adaptive hybrid genetic algorithm and applied to support vector machine model parameter selection research. Simulation results show that the selected parameters in the algorithm used in SVM model parameter selection than traditional genetic algorithm, cross-validation and grid search algorithm selected parameters should improve the recognition accuracy of SVM. Four, according to the special significance of the parameters ε and v ε-SVR and v-SVR, outlier detection method based on the ε-SVR regression analysis and outlier detection method based on the v-SVR regression analysis. The experimental results show that the proposed outlier detection method based on ε-SVR regression analysis and regression analysis process can accurately and effectively detect outlier detection method based on the regression analysis of the v-SVR in isolated points.

Related Dissertations

  1. Research on Automatic Detection Algorithm for Substructure Distress of Highway Pavement Based on SVM,U418.6
  2. Research on Autamatic Music Structrue Analysis,TN912.3
  3. Research on Transductive Support Vector Machine and Its Application in Image Retrieval,TP391.41
  4. Fault Diagnosis Method Based on Support Vector Machine,TP18
  5. Process Support Vector Machine and Its Application to Satellite Thermal Equilibrium Temperature Prediction,TP183
  6. Research for Infrared Image Target Identification and Tracking Technology,TP391.41
  7. Study on the Road Condition Monitoring Based on Vehicular 3D Acceleration Sensor,TP274
  8. Research of Diagnosing Cucumber Diseases Based on Hyperspectral Imaging,S436.421
  9. The Research on Intrusion Detection System Based on Machine Learning,TP393.08
  10. Research on Improved K Neighbor Support Vector Machine Algorithm Faced Text Classification,TP391.1
  11. Research on Face Recognition Based on AdaBoost Algorithm,TP391.41
  12. Research on Feature Extraction, Selection and Classification Algorithms for Pulmonary CAD,TP391.41
  13. Research on Subimage Selection and Mathching Method for Synthetic Aperture Radar(SAR) Target Recognition,TN957.52
  14. Research of Facial Expression Recognition Algorithm,TP391.41
  15. Fundus Image Segmentation Based on SVM and Template Matching,TP391.41
  16. Research and Realization of License Plate Character Recognition Algorithm Based on SVM,TP391.41
  17. Cloud Classification Based on Geostationary Meteorological Satellite Imagery,TP391.41
  18. The Studies on Some Improvements of the GA and Their Applications in SVM,TP18
  19. The Research and Implementation of the Fault Management System for Triple Play,TP315
  20. Modulation Classification Algorithms of Digital Communication Signals,TN914.3
  21. Computer virus detection method,TP309.5

CLC: > Industrial Technology > Automation technology,computer technology > Automated basic theory > Artificial intelligence theory
© 2012 www.DissertationTopic.Net  Mobile