Optimization Simulation of Network Retrieval Model Based on Big Data Algorithm

The traditional keyword-based information retrieval service can only return the results matching keywords according to the user’s query conditions, and the retrieval mechanism is relatively simple, leaving out a lot of information with similar content but different expressions. This paper proposes and implements an optimization method of network retrieval model based on big data algorithm. The model expanded by term similarity includes arcs that simulate the direct relationship between terms, which mainly expand the term subnet. According to the quantitative synonym relationship between terms, an extended BN (Bayesian network) retrieval model is established, and the retrieval results are presented to users in descending order of document relevance, thus realizing semantic-based information retrieval to some extent. The simulation results show that the average precision/recall ratio of the proposed retrieval strategy is higher than that of the traditional retrieval model, and the average precision ratio can reach 66.37%. Secondly, because the document is indexed offline, it is also better than the traditional retrieval model retrieval strategy in time consumption. Experiments show that compared with the traditional retrieval model, the proposed model has better retrieval performance. The model in this paper can expand the retrieval of related information without deviating from the user’s retrieval goal, and the ranking of related documents is more reasonable, and the quality of retrieval results is further improved.