1 / 20

On Utillizing LVQ3-Type Algorithms to Enhance Prototype Reduction Schemes

On Utillizing LVQ3-Type Algorithms to Enhance Prototype Reduction Schemes. Sang-Woon Kim and B. John Oommen* Myongji University, Carleton University*. Outline of the Study. Introduction Overview of the Prototype Reduction Schemes The Proposed Reduction Method Experiments & Discussions

metea
Télécharger la présentation

On Utillizing LVQ3-Type Algorithms to Enhance Prototype Reduction Schemes

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. On Utillizing LVQ3-Type Algorithms to Enhance Prototype Reduction Schemes Sang-Woon Kim and B. John Oommen* Myongji University, Carleton University*

  2. Outline of the Study • Introduction • Overview of the Prototype Reduction Schemes • The Proposed Reduction Method • Experiments & Discussions • Conclusions Workshop on PRIS’ 2002

  3. Introduction (1) • The Nearest Neighbor (NN) Classifier : • A widely used classifier, which is simple and yet one of the most efficient classification rules in practice. • However, its application often suffers from the computational complexity caused by the huge amount of information. Workshop on PRIS’ 2002

  4. Introduction (2) • Solving strategies to the problem : • Reducing the size of the design set without sacrificing the performance. • Accelerating the speed of computation by eliminating the necessity of calculating many distances. • Increasing the accuracy of the classifiers designed with limited samples. Workshop on PRIS’ 2002

  5. Motivation of the Study • In NN classifications, prototypes near the boundary play more important roles. • The prototypes need to be moved or adjusted towards the classification boundary. • The proposed approach is based on this philosophy, namely that of creating and adjusting. Workshop on PRIS’ 2002

  6. Prototype Reduction Schemes- Conventional Approaches - • The Condensed Nearest Neighbor (CNN) : • The RNN, SNN, ENN, mCNN rules • The Prototypes for Nearest Neighbor (PNN) classifiers • The Vector Quantization (VQ) & Bootstrap (BT) techniques • The Support Vector Machines (SVM) Workshop on PRIS’ 2002

  7. A Graphical Example (PNN) Workshop on PRIS’ 2002

  8. LVQ3 Algorithm • An improved LVQ algorithm: • Learning Parameters : • Initial vectors • Learning rates : • Iteration numbers • Training Set = Placement + Optimizing: Workshop on PRIS’ 2002

  9. Support Vector Machines (SVM) • The SVM has a capability of extracting vectors which support the boundary between two classes, and they can satisfactorily represent the global distribution structure. Workshop on PRIS’ 2002

  10. Extension by Kernels Workshop on PRIS’ 2002

  11. The Proposed Method • First, the CNN, PNN, VQ, SVM are employed to select initial prototype vectors. • Next, an LVQ3-type learning is performed to adjust the prototypes: • Perform the LVQ3 with Tip to select w • Perform the LVQ3 with Tip to select e • Repeat the above steps to obtain the best w* and e* • Finally, determine the best prototypes by invoking the learning n times with Tip and Tio. Workshop on PRIS’ 2002

  12. Experiments • The proposed method is tested with artificial and real benchmark design data sets, and compared with the conventional methods. • The one-against-all NN classifier is designed. • Benchmark data sets : Workshop on PRIS’ 2002

  13. Workshop on PRIS’ 2002

  14. Workshop on PRIS’ 2002

  15. Experimental Results (3) Workshop on PRIS’ 2002

  16. Experimental Results (4) Workshop on PRIS’ 2002

  17. Data Compression Rates Workshop on PRIS’ 2002

  18. Classification Error Rates (%)- Before Adjusting - Workshop on PRIS’ 2002

  19. Classification Error Rates (%)- After Adjusting with LVQ3 - Workshop on PRIS’ 2002

  20. Conclusions • The method provides a principled way of choosing prototype vectors for designing NN classifiers. • The performance of a classifier trained with the method is better than that of the CNN, PNN, VQ, and SVM classifier. • The future work is to expand this study into large data set problems such as data mining and text categorization. Workshop on PRIS’ 2002

More Related