Name
Abstract | ||
|---|---|---|
In this paper, we propose KNC algorithm for combining KNN algorithm and other three classifiers (C4.5 algorithm, Naive Bayes classifier and SVM) based on their classification capabilities on different types of instances. According to labels of instances and their K nearest neighbors, we divide instances into three types, S-, DS- and D-type. The classification capabilities of KNN algorithm on S-type instances are the best, while ones on D-type and DS-type are usually worse than other three classifiers. KNC algorithm uses KNN algorithm to classify S- and DS-type instances, and uses other classifiers to classify D-type instances. KNC algorithm utilizes classification capability of KNN algorithm on S-type instances and utilizes classification capabilities of other three classifiers on D-type instances. Experimental results on 20 UCI data sets demonstrate utility and feasibility of KNC algorithm. © 2010 IEEE. |
Year | DOI | Venue |
|---|---|---|
2010 | 10.1109/COGINF.2010.5599804 | IEEE ICCI |
Keywords | Field | DocType |
c4.5 algorithm,ensemble learning,knn algorithm,naive bayes classifier,svm,support vector machines,artificial neural networks,accuracy,prediction algorithms,niobium,classification algorithms,c4 5 algorithm,k nearest neighbor | k-nearest neighbors algorithm,Data set,Naive Bayes classifier,Pattern recognition,Computer science,Support vector machine,C4.5 algorithm,Artificial intelligence,Artificial neural network,Statistical classification,Ensemble learning,Machine learning | Conference |
Volume | Issue | Citations |
null | null | 1 |
PageRank | References | Authors |
0.36 | 4 | 2 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Zhiyong Yan | 1 | 13 | 2.55 |
| Congfu Xu | 2 | 131 | 15.71 |