| With the rapid development of the economy and the continuous improvement of people's quality of life,tourism has become one of the fastest growing industries.In terms of tourism and leisure,people began to inquire about travel information through the Internet and select the information they like.The increase in users has led to an exponential increase in the amount of data on the Internet.The increase in the amount of data has led to a large amount of time for users to screen their favorite attractions of interest,making it difficult to select a travel route that suits them.The recommendation system helps users select their favorite travel routes from big data.In the field of travel recommendation,the traditional recommendation algorithm has achieved good results,due to the lack of data,the cold start and data sparseness problems and the neglect of the semantic problems hidden in the travel track,the low recommendation accuracy remains unresolved.This paper uses sequence learning-based tourist attraction recommendations to solve these problems.The method provides personalized tourist attraction recommendation for tourists by analyzing the tourist trajectory information of tourists,aiming at solving in the traditional recommendation algorithm the difficulty of lack of data-cold start and data sparsity problem,neglecting the implicit semantic problem in the travel track,with the low recommendation accuracy.The research content of this paper is as follows:This paper first proposes a recommendation method for based on the gated recurrent unit neural network.It is designed to solve the problem of starting difficulties caused by the lack of data in the traditional recommendation algorithm-cold start,sparse data,neglecting the problem of high-level semantics in the travel track and low recommendation accuracy.Through the gated recurrent unit network,the whole travel trajectory is modeled,and multiple travel trajectories are simultaneously trained.At the same time,the same location in other travel trajectories is used as a negative sample to train,reduce the amount of calculation and time of the negative example,use the historical trajectory of tourists to provide visitors with personalized tourist attractions.Experiments on real tourism datasets show that the methods used in this paper are significantly more effective than the baseline methods widely used in recommendation systems.Based on the recommendation method of the gated recurrent unit neural network based on the attraction recommendation method,an optimization method model based on the maximum negative sample is proposed.On the basis of the original model,the effect ofthe effective negative sample on the recommendation effect is considered.The other negative samples score are compared with the largest negative sample score to get the weight.The higher the score of the negative sample,the greater the weight obtained.The higher the weight of the negative sample,the greater the influence on the recommendation result.The negative weight of the negative sample with low score has little effect on the recommendation effect.The negative sample is added on the original basis and the negative sample with the highest score is compared with all the negative sample and the weight assignment is performed.Improve in this way,recommend personalized tourist attractions for each visitor. |
