Chinese Subjective Sentence Extraction Based on Dictionary and Combination Classifiers

Abstract

For extracting of Chinese subjective sentence, this paper proposes a new dictionary-based extraction method and a novel classifier combination strategy. For the first method, we use the training data to score the subjective dictionary, which was composed of indicative verb, indicative adverbs, sentiment words, interjection and punctuation. Then we use the dictionary to score the test data, and filter the sentences by setting a reasonable threshold. New classifier combination strategies base on the maximum error correction capability. To enhance the accuracy, the method improves the traditional single error correction and achieves the dual error correction both in positive and negative classes. Experimental results show that the two methods are effective .And the final results show that the combination of two ways achieves a satisfactory subjective sentence extraction performance.