In the preceding years the mode of answering to the query is used as a by hand. Due to enlarge in use of internet this method shifted to internet. This project is mainly implemented to provide question and answer type in the form of internet. This mode is easy to conduct the large amount of free online text. Based on VerbNet frames this project is implemented. It elaborates the syntactic, thematic and semantic information. This is because to filter out unmatched sentences in semantic level and to elaborate answer portion from the respond sentence. VerbNet is applied to our project to detect the verb frames in question and candidate sentence. Using these method syntactic and thematic information as well as semantic information can be therefore obtained.

            Regarding the background information the user raises the question to the system then the system analyzes and raises two are more queries to the user, after getting the feedback from the user the system answers to the question. Charles J.Fillmore has introduced the case grammar and it defines the sentence meaning. VerNet is a hierarchical domain-independent which contains the syntactic information. It contains a 5257 verb senses and 274 frame classes. The way of processing to the query is of four types. Syntactic analysis and preprocessing, word order adjustment, Tense and voice unification, VerbNet frame detection.

            Our approach to this project is verb frames in question and candidate sentences are detected and we will get the thematic information. When the data present in the answer matches to the question chunks it detects the answer. The main theme of this project is to filter not matching sentences and to rank the correct answer as right.