Research On Context-aware Recommendation Technology Based On Distance Feature And Time Effect

With the advent of the Web2.0 era,the rapid development of network technology and the popularity of mobile devices,the amount of data generated in the Internet has grown exponentially,and users have been drowned in the vast ocean of data.In order to solve this problem,the recommendation system came into being.It can filter out most of the spam information for users,and saves the user's time while solving the information overload problem to some extent.In many scenarios,contextual information is an important factor in determining user interest.Context-aware recommendation systems can improve the accuracy of predictions in many scenarios and are becoming a research hotspot.Distance and time are the two most common factors in designing context-aware recommendation systems.Therefore,this paper will improve the recommendation algorithm and compare the results in order to improve the accuracy and recall rate of the algorithm,and discuss how to improve the accuracy and recall rate of the algorithm by considering only the physical information in the distance and ignoring the cultural differences in the distance and the time decay effect in the time factor and ignoring the time memory effect.Improving the accuracy of recommendation results in case of missing.(1)Context-aware recommendation technique based on feature distance.In the current study,the distance factor generally refers to the geographical or physical distance that can be directly measured.However,in the recommendation system,each decision made by the user is not only related to his geographical location,but also closely related to the generalized distance characteristics of the family environment,economic status,religious beliefs and values,so only narrow sense is considered.The recommended system of distance characteristics has certain limitations.This paper quantifies the abstract concept of the generalized distance feature,introduces Hofstede's cultural distance model(a model for measuring cultural differences and value orientations in different countries)in the collaborative filtering recommendation system,and proposes a collaborative filtering recommendation algorithm based on feature distance and predict the values of some items with missing cultural distance values.Algorithm and predict the distance eigenvalues of some countries that lack cultural distance values.Finally,experimental verification is performed on the real data set.The experimental results show that the proposed algorithm can improve the accuracy by 9%compared with the traditional recommendation method.(2)Context-aware recommendation technology based on time effect.The traditional time context-aware recommendation system only considers the time decay effect,and believes that the correlation between users will only decrease with time,ignoring the specific memory and enhancement of people under certain time conditions.The correlation between the two.In order to better analyze the user's intention and produce more accurate recommendation results,this paper discusses the influence of time effect on the recommendation system,proposes the concept of time memory effect,and combines it with the traditional time decay effect to propose a new time effect function..Then,according to the new time effect function,the weight of the user similarity is changed and the predicted score of the recommended item is affected,and finally the recommendation result is generated according to the predicted score.Experimental results based on large-scale actual data show that the context-aware context-aware recommendation algorithm can improve the accuracy by 7%compared with the traditional recommendation method.