A CONTEXT SENSITIVE AND PERSONALIZED QUERY AUTOCOMPLETION TECHNIQUE

Abstract

Query Autocompletion is a leading attribute of Search Engines which makes the user’s search experience better by predicting the query. QAC methods suggest query suggestions to users, after they enter some of the keystrokes in the search engine. This is done by predicting the query using past query logs and other trends. Current QAC methods use the Most Popular Completions as the suggestion results. Context and Personalized techniques are proposed already but they are used separately. The present methods being incorporated are the location and past searches sensitive QAC. In this proposed work of thesis, we will talk about a hybrid technique by combining both the context sensitive, trending and personalized suggestions. The improvements which are made in the base paper are that a new approach can be proposed by combining the three techniques to create a hybrid technique. It intends to incorporate three major research works: Time sensitive (based on time series and trends), Context Sensitive (based on recent searches done) and Personalized (based on gender, location and age-group) query auto completion. Thus an algorithm that considers all these parameters will be better at predicting the user query. The results predicted are better in reducing the user keystrokes during the search and also reduces the searching time, and also enhances the reliability of the search engine. Further improvements can be done by extracting the user’s browsing history to determine keywords, interests and other user-specific data for enhancing the result predictions.