Automatic Identification of Query Intent by Using Multiple Features
Jiang Zong-l · Computer and Information Technology · 2015
In order to improve the quality of ranking result from search engine, increasing attentions are paid to identify the underlying user intents. In this paper, we identify user's query intent with the multidimensional feature extracted from the queries based on sessions. Then classify by using SVM. Experiment result show that by combing multiple features help identify user intent, classification accuracy is up to 80% in our test data set manually annotated.