Research On Data Fusion In Information Retrieval By Using Intelligent Optimization Methods

With the development of modern information technology,how to help users get the information they need quickly and accurately from a large amount of information on the Internet is the most important problem in information retrieval.The data fusion technology in information retrieval combines the retrieval results from different retrieval systems so as to get a new retrieval result.The previous research shows that the data fusion technology can effectively improve the performance of retrieval results.In this research,we mainly study the linear combination method in data fusion technology,and focus on how to solve the weight assignment problem of the linear combination method by using the intelligent optimization algorithm.The main work of this dissertation is as follows:(1)In this dissertation,we discuss the weight allocation strategy based on the differential evolution algorithm and the particle swarm optimization algorithm.On the basis of the above two optimization algorithms,we present a weight allocation strategy of self-adaptive alternating optimization algorithm.In this method,the weight of the system is optimized by using the adaptive probability between the differential evolution algorithm and the particle swarm optimization algorithm.This is the first time for this method to be applied to data fusion.(2)In order to test the effectiveness of the above fusion methods,we used the TREC 2004 Robust Task data sets for the experiments.The experimental results show that all three methods are able to improve fusion performance and particularly the last one is the best among them.(3)We also compare the time complexity of the three methods for training weights.Among them,the particle swarm optimization algorithm is the fastest,which is followed by the mixed method of differential evolution and particle swarm optimization,while the differential evolution algorithm takes the longest time.In summary,this dissertation gives a brief description of the existing data fusionmethods and proposes a new weight allocation strategy for the linear combination method by applying several intelligent optimization algorithms.The effectiveness and efficiency of these fusion methods are compared through experiments.Considering time and performance together,the experimental results show that the mixed method is the best.