Improved Sentence Similarity Algorithm based on VSM and its application in Question Answering System

In the FAQ-based Chinese Question Answering System, the most critical issue is how to calculate the similarity between the user questions and the questions in the FAQ. The traditional VSM-based Sentence Similarity Algorithm usually regards word as the basic linguistic unit of sentences and mainly considers the statistical information of words in questions, but doesn´t take the word importance in the professional field and the semantic information of words into account. For these reasons, this paper proposes an Improved Sentence Similarity Algorithm Based on VSM, regarding notion as the basic linguistic unit of sentences, through conceptually abstracting and professionally classifying to improve the performance of Sentence Similarity Algorithm. Testing in Chinese FAQ system of specific areas, experimental result shows that the performance of the improved algorithm is superior to the traditional VSM-based sentence similarity algorithm evidently.