Multi-feature-fused matrix decomposition interest point recommendation method and implementation system thereof

Abstract

The invention discloses a multi-feature-fused matrix decomposition interest point recommendation method and an implementation system thereof. Firstly, according to the influence of friends and non-friends in the social relation on user decisions, the personalized distribution of user sign-in is calculated by using a self-adaptive bandwidth kernel density method in combination with user scores, andthe correlation between interest points is obtained; and then, because the sequence output by the Bi-LSTM has the characteristics of word semantics, syntax between the front and back of the word sequence and other hidden information, and the CNN is skilled to capture significant characteristics from a series of characteristics, the Bi-LSTM and the CNN are superposed to form a new deep neural network, thereby learning the potential characteristics of the user and the interest point. Finally, the social contact, the geographic position, the classification preference and the potential features are fused through a probability matrix method, the personalized preference of the user is predicted, and therefore the purpose of personalized recommendation is achieved.

Description

technical field [0001] The invention relates to the technical field of information recommendation, in particular to a method for recommending points of interest based on matrix decomposition based on fusion of multiple features and an implementation system thereof. Background technique [0002] With the continuous development of smartphones and smart devices, applications based on Location Based Social Networks (LBSN) (such as Yelp, Foursquare, Streetside, etc.) attract more and more users to share their check-in status, location, etc. and related comments. Location-based social network POI recommendation is to provide users with the most attractive and relevant POIs (such as hotels, restaurants, scenic spots, etc.) by filtering massive information in social networks, and reduce the negative impact of information load. Thus promoting the continuous development of POI recommendation in the era of big data. [0003] In recent years, Point of Interest (POI) recommendation has...

AI Technical Summary

Problems solved by technology

However, individual models have more or less their own shortcomings, which are often ignored

Therefore, using a single model to learn the content representation of the review text may forget or ignore some of the content in the text, resulting in biased recommendation results

Images