A Chinese Classifier Research for Query Intention and Non-Query Intention

摘要

Previous research to improve the performance of Internet search engines has focused on classifying questions, sentences and user-goals but not the classification of sentences and phrases based on query intention and non-query intention. This paper investigates a classification system of query intention and non-query intention of sentences and phrases by firstly analyzing previous work and based on this we move on to designing an implementation framework model that improves the efficiency and effectiveness of using search engines. Testing was carried out and comparisons analyzed between the known Liner, Polynomial, RBF and Sigmoid categorizations and our new framework under strict laboratory conditions. The test results show that the accuracy of classifier classification is improved by up to 2.1% and recall is improved by up to 2%.