An evaluation of statistical approaches to text categorization-英文文献.pdfVIP

Information Retrieval 1, 69–90 (1999) c 1999 Kluwer Academic Publishers. Manufactured in The Netherlands. An Evaluation of Statistical Approaches to Text Categorization YIMING YANG yiming@ School of Computer Science, Carnegie Mellon University, Pittsburgh, PA 15213-3702, USA Received October 28, 1997; Revised May 13, 1998; Accepted July 27, 1998 Abstract. This paper focuses on a comparative evaluation of a wide-range of text categorization methods, in- cluding previously published results on the Reuters corpus and new results of additional experiments. A controlled study using three classifiers, kNN, LLSF and WORD, was conducted to examine the impact of configuration vari- ations in five versions of Reuters on the observed performance of classifiers. Analysis and empirical evidence suggest that the evaluation results on some versions of Reuters were significantly affected by the inclusion of a large portion of unlabelled documents, mading those results difficult to interpret and leading to considerable con- fusions in the literature. Using the results evaluated on the other versions of Reuters which exclude the unlabelled documents, the performance of twelve methods are compared directly or indirectly. For indirect compararions, kNN, LLSF and WORD were used as baselines, since they were evaluated on all versions of Reuters that ex- clude the unlabelled documents. As a global observation, kNN, LLSF and a neural network method had the best performance; except for a Naive Bayes approach, the other learning algorithms also performed relatively well. Keywords: text categorization, statistical learning algorithms, comparative study, evaluation 1. Introduction Text categorization (TC) is the problem