Dissertation > Excellent graduate degree dissertation topics show

Research on Support Vector Machine Solving Large-scale Data Set

Author: JiXiuHao
Tutor: ZhangZuoGang
School: Liaoning Technical University
Course: Applied Mathematics
Keywords: SVM Rough Set RS-SVDD Attribute Reduction SVDD
CLC: TP18
Type: Master's thesis
Year: 2011
Downloads: 47
Quote: 0
Read: Download Dissertation

Abstract


Support Vector Machines (SVMs) algorithms based on the foundations of Statistical Learning Theory study on the classification and prediction for small sample data. Based on complete mathematical theory SVMs bring on a good generalization performance by using of the structure risk minimization principle.So it become a hotspot in machine learning and artifical intelligence. But there are several shortcomings, such as the selection of modle, the classification of multi-class and the training of lagre scale samples.In this thesis, we analyzed the theory of the svm algorithms, and described the process of iterative and training for different SVM algorithm models. In order to implement the better ability of classification and regression for lager scale data, the improvement methods are proposed about the sample data reduction, knowledge extraction and algorithm training. The simulation results show that these methods can effectively solve the problem of large-scale sample data set.The main research in this paper can be classed as follows:Firstly, the thesis systematically analyzed the basic theory of support vector machines and the mathematical models, described the process of transforming the training problem to a convex quadratic programming with constraints problem, summarized the SVM generalization performance, and compared some of the popular fast training algorithms in detail.Secondly, the strategies of attribution reduction and knowledge rules extraction based on rough set are studied in detail. The different algorithms of rough set are summarized. The least squares support vector machine algorithm model based on rough set is proposed. The performance of the algorithm is good proved from the classification and regression. Based on the sample data set collected from MATLAB 802.11a, the models of the channel classification and parameter regression can be established with least squares support vector machine. The channel selection and parameters configuration are determined by the models. The effectiveness of the algorithm can be proved by the several simulation experiments.Thirdly, Support Vector Data Description (SVDD) algorithm is studied deeply. In order to reduce the computational cost, the algorithm of Support Vector Data Description based on Rough Set (RS-SVDD) is proposed. The large-scale sample data would be reduced by rough sets and SVDD algorithms, and the classification model can be established. In the aspect of classification accuracy and the training speed, RS-SVDD algorithm is better than SVDD by evaluating the simulation results.

Related Dissertations

  1. Soft Sensor of Naphtha Dry Point on Support Vector Machines Regression,TE622.1
  2. The Research of the Fault Diagnoses Algorithm for the Liquid Rocket Engine Testing Bed Based on PCA-SVM,V433.9
  3. ISAR Imaging Simulation of Space Targets and Target Recognition Based on ISAR Images,TN957.52
  4. Research on Feature Extraction and Classification of Pulse Waveform for Cholecystitis and Nephrotic Syndrome Diagnosis,TP391.41
  5. The Research on Paper Currency Classification Method Based on Harr-Like Feature and Minimal Ball Including Samples,TP391.41
  6. Fault Diagnosis Method Based on Support Vector Machine,TP18
  7. Research on Clustering Algorithm Based on Genetic Algorithm and Rough Set Theory,TP18
  8. Based on Rough Set of Urban Areas When Traffic Green Control System Research,TP18
  9. Research of Orange Quality Classification Technology Based on Computer Vision,TP391.41
  10. Calculation of Knowledge Granulation and Study of Its Application in Attribute Reduction,TP18
  11. Study on Automatic Classification Method for Remotely Sensed Imagery by Incorporating Spatial-Spectral Features,TP751
  12. The Applicability of Data Mining Technology for Environmental Protection Management System,TP311.13
  13. Researching of Scheduling and Control Co-design Based on Variable Sampling Period in Networked Control System,TP273.5
  14. The Research on Electrode 3D Model Classification and Retrieval Based on SVM and Shape Features,TP391.41
  15. Study on the Decision Tree Classification Algorithm and Its Application Based on Rough Set Theory,TP18
  16. Based on the core set of examples of attribute reduction method,O159
  17. Expression of face recognition algorithm based on sparse research,TP391.41
  18. Based on PCA and SVM automotive coating line electromechanical equipment intelligent diagnosis,TH165.3
  19. The Research on the Influence Factors and Control Countermeasures of the Implicit Cost in Construction Stage of Construction Project,F285
  20. Study on the Engineering Properties of Loess and the Stability of Loess Slope Along a Railway,U212.22
  21. Application of QSAR and Molcular Docking Studies in Medicinal Analytical Chemistry,R917

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