Research On Keyword Search Results Diversification

Web search engines and databases usually sort query results based on relevance.However,recently researchers are more and more inclined to pay attention to diversity,because it can help improve user satisfaction.Therefore,there are a lot of applications of diversification in many fields.Take the service composition in the software system development process as an example.We introduce the application of diversification in query and retrieval.We propose the problem of diversified service composition in the process of software development,and provide two algorithms,All-Then-Diversify algorithm and Pop-AndDiversify algorithm,to solve the service composition retrieval problem that it is too slow to diversify results on graph.Finally,we verify that our algorithms can effectively reduce query time ensuring diversity in the real data set.Then,we delve into the characteristics of diversity,put forward the concepts of local diversity and global diversity,and construct a three-criteria optimization problem.In the process of solving the problem,two properties are discovered and verified.Based on these two properties,a heuristic greedy search algorithm is proposed.And,we added a vertex replacement algorithm to optimize the results of the greedy search algorithm.We have verified through experiments that our algorithm can effectively produce more diverse results.With the development of diversity,hierarchical diversity has been proven to be able to better diversify query results.Therefore,we propose a hierarchical and diverse framework.Through the hierarchical clustering algorithm,the preliminary candidate results are constructed into a hierarchical tree.According to the hierarchical tree,we can calculate the diversity of results.Finally,we verify the effectiveness and efficiency of our model through experiments.