Research And Application Of Exploratory Search Technology For The Gestural Interaction

With the rapid development of human-computer interactive technology,the natural human-computer interactive technology represented by gestural interaction is widely populated by users.The traditional information search technology can not fully meet the exploratory search needs of gestural interaction,which requires gestural interaction process to predict possible query intent and provide a variety of query results' list for users to find more useful information,so this thesis study the gestural query intent modeling method and diverse ranking algorithm for the gestural interaction requirements.The main work is as follows:(1)This thesis analyzes the problems brought by the rapid development of gestural interaction and clarifies the significance of the research.The thesis investigates the research and application of exploratory search for gestural interaction.(2)In order to solve the problem of the useless gestures,reduce the burden of the query and help the user to obtain the required content,a gestural query intent model(GQIM)based on Hidden Markov Feature is proposed and predicts the optimal query intention sequence corresponding to the touch interaction event based on the Viterbi theory.The simulation results show that the proposed method has a certain effect on improving user satisfaction.(3)In order to improve the balance between the diversity and the correlation about the existing query result sets and improve the fluency of gestural interaction,this thesis studies the ranking of query result sets and a new diverse ranking algorithm based on the minimal independent dominating set(MIDS-DR)is proposed.The problem of diverse subsets selected is transformed into minimal independent dominating set solving the problem of undirected weighted graphs,so as to balance the diversification and relativity of query results.In the process of solving,the concept of abandoned subset reduces the comparison of the distance between redundant vertex pairs and speeds up the algorithm.The simulation results show that the proposed algorithm can improve the diversified performance and query processing efficiency.(4)Aiming at the fluency and satisfaction of user gestural interaction,a Wikipedia search system for the gestural interaction is designed and implemented.The functions of exploratory search keywords,manipulating keyword distributed graph and viewing the diverse ranking lists are realized,and use the corresponding tested tool to test the search system.