Short Text Categorization Based on Instance Selection Learning
Yan ZHAN, Hao CHEN, Guochun ZHANG
College of Mathematics and Information Science, Hebei University, Baoding, 071002, CHINA
Abstract: Text Categorization is an important component in many information organization and information management tasks. In short text classification problem, which is as a branch of text classification, there will be too many instances which need much computation time and memory requirement. This paper proposes an instance selection learning method which can reduce the instance numbers in K-NN classification algorithm. The experiments also compared the learning algorithm with existing reducing samples algorithms such as Condensed Nearest Neighbor, Selective Nearest Neighbor, Reduced Nearest Neighbor Rule, Edited Nearest Neighbor Rule in Short Text Categorization.
Keywords: Short Text Categorization; Instance Selection Learning; K-NN Classification