Research On Expert Recommendation Of Community Of Questions Answering Based On Sailfish Optimization Algorithm

With the rapid development of mobile Internet,community of questions answering has gradually become an important channel for people to obtain information.More and more users enter into the community of questions answering platform,and a large number of new questions are constantly generated in the community.When new problems continue to arise,and the problem of historical accumulation cannot be solved,it is difficult for users to quickly obtain information in the community of questions answering to meet their knowledge needs.In order to solve this problem,it is necessary to recommend the right problem to the right experts,so that the problem can be solved quickly and with high quality.At present,expert recommendation in community of questions answering is to recommend questions to users one by one for answers.This paper proposes a method of batch recommendation to optimize the use of expert resources.Based on the multi-objective optimization algorithm,it aims to find a set of questions with large coverage of the original knowledge demand from a large number of new questions and recommend them to experts.When consuming less expert resources,it makes the recommended questions as much as possible There is also a high probability that the problem set will be answered by experts.The main work is as follows:Firstly,the BTM model is used to model the problem text.Based on the continuous single objective sailfish optimization algorithm,the discrete sailfish optimization clustering method is proposed to cluster the historical questions answered by experts,and different question fields are obtained,which provides the basis for the calculation of question field and expert field.Then,the similarity between the new problem and each domain is calculated,and the domain correlation matrix of the problem is obtained.At the same time,in order to measure the level of experts’ professional knowledge in various fields,the latest response time,cumulative response frequency and average praise number of experts in various fields are calculated to get the ranking of experts in each field.Finally,the process of expert recommendation is transformed into a multi-objective problem.Aiming at the low coverage of the recommended problem to the original problem set,the low consumption of expert resources,and the high probability of being successfully answered,a binary multi-objective sailfish optimization algorithm with genetic algorithm is constructed to solve the problem.In order to alleviate the problem of information redundancy,knowledge demand and supply mismatch in community of questions answering,so as to improve the satisfaction of community of questions answering,promote knowledge sharing and maintain the healthy development of the community.In this paper,we use the real data in the Zhihu community of questions answering to carry out the experiment.In the way of contrast experiment,we verify it from two aspects:other recommendation strategies and other swarm intelligence optimization algorithms.The experimental results show that the proposed method is feasible and practical.