Personal Location Recommendation Based On Information Entropy And Trust Mechanism

Personalized location recommendations have become a research hotspot in the field of recommendation in recent years.Although some scholars have done related researches on location recommendations and obtained corresponding research achievements,there are still some problems: First,with the user location check-in data that contains users' preference,most of the methods directly construct the scoring matrix with regarding it as explicit data.Second,when constructing the user model with the check-in data,most of the methods adopt Gaussian distribution.However,because of the skewness of the check-in data,the Gaussian distribution is not suitable for modeling it.Third,though some scholars realized that geographical factors have different influences on different users,they did not realize that different analysis scales should be used in areas with different sign-in densities.Therefore,based on the existing work,this thesis studies the personalized location recommendation algorithm based on information entropy and trust mechanism with adopting the principle of information entropy and trust mechanism.The main work and contributions of this thesis are as follows:(1)Being aimed at the serious inclination of the check-in data,we define the location information to normalized the check-in data that can reflect the user's preference with drawing lessons from the concept and principle of information entropy.Based on the user location information data,we improve a personalized location recommendation algorithm based on information entropy.The Poisson distribution integrated with location importan is used to model the check-in data,to improve performance of the recommendation algorithm.(2)To alleviate the cold-start problem,we improve a personalized recommendation algorithm based on trust mechanism.Finally,using the cascading hybrid recommendation method to combine the personalized location recommendation algorithm based on information entropy and the personalized location recommendation algorithm based on trust mechanism to the personalized location recommendation based on information and trust mechanism,and integrated with the geographic influence to give a final list of recommended placed for Improving Recommendation Performance of Algorithms.