{"title":"用模糊积分技术改进SOM分类器","authors":"A. Jirayusakul","doi":"10.1109/ICTKE.2012.6152395","DOIUrl":null,"url":null,"abstract":"As Self-organizing map (SOM) neural network is implemented as a pattern classifier. According to the decision process of the SOM classifier, the traditional technique, called the winner-take-all, is employed to search the final class of an unknown input. In practice, some prototypes on the SOM classifier might not be representatives of purity class regions. Hence, the decision process of the SOM requires information about both the winner prototype and its neighbors to improve an accuracy rate. In this paper, the Fuzzy Integral decision technique is applied to aggregate information about the winner prototype and its neighbors for determining the final class of an unknown input. The experimental results of the UCI datasets showed that the proposed decision technique could improve accuracy rates better than the traditional technique.","PeriodicalId":235347,"journal":{"name":"2011 Ninth International Conference on ICT and Knowledge Engineering","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2012-02-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":"{\"title\":\"Improve the SOM classifier with the Fuzzy Integral technique\",\"authors\":\"A. Jirayusakul\",\"doi\":\"10.1109/ICTKE.2012.6152395\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"As Self-organizing map (SOM) neural network is implemented as a pattern classifier. According to the decision process of the SOM classifier, the traditional technique, called the winner-take-all, is employed to search the final class of an unknown input. In practice, some prototypes on the SOM classifier might not be representatives of purity class regions. Hence, the decision process of the SOM requires information about both the winner prototype and its neighbors to improve an accuracy rate. In this paper, the Fuzzy Integral decision technique is applied to aggregate information about the winner prototype and its neighbors for determining the final class of an unknown input. The experimental results of the UCI datasets showed that the proposed decision technique could improve accuracy rates better than the traditional technique.\",\"PeriodicalId\":235347,\"journal\":{\"name\":\"2011 Ninth International Conference on ICT and Knowledge Engineering\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2012-02-16\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"2\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2011 Ninth International Conference on ICT and Knowledge Engineering\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICTKE.2012.6152395\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2011 Ninth International Conference on ICT and Knowledge Engineering","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICTKE.2012.6152395","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Improve the SOM classifier with the Fuzzy Integral technique
As Self-organizing map (SOM) neural network is implemented as a pattern classifier. According to the decision process of the SOM classifier, the traditional technique, called the winner-take-all, is employed to search the final class of an unknown input. In practice, some prototypes on the SOM classifier might not be representatives of purity class regions. Hence, the decision process of the SOM requires information about both the winner prototype and its neighbors to improve an accuracy rate. In this paper, the Fuzzy Integral decision technique is applied to aggregate information about the winner prototype and its neighbors for determining the final class of an unknown input. The experimental results of the UCI datasets showed that the proposed decision technique could improve accuracy rates better than the traditional technique.