Interest point recommendation method based on explicit feature and implicit feature fusion

Abstract

The invention discloses a point-of-interest recommendation method based on explicit feature and implicit feature fusion, which comprises the following steps: firstly, through data preprocessing, obtaining the sign-in time and region of a user, the region of a point-of-interest and the time information when the point-of-interest is accessed, and then extracting explicit vectors of the user and the point-of-interest from the region of the user and the point-of-interest; the invention further provides an FGMF model, the GMF model is improved, the obtained explicit feature vector and the n-dimensional implicit feature vector are input into the FGMF model together to be pre-trained, then the result is input into a deep neural network to be trained, the score of each user on the interest point is predicted, then the scores are sorted, and the interest point is obtained. And recommending and generating a top-k recommendation list for each user according to the score. According to the method, the influence of the explicit features and the implicit features on interest point access is considered at the same time, and the recommendation accuracy is improved.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to a method and system for recommending points of interest based on a deep neural network. Background technique [0002] With the continuous development of urban rail transit, our outdoor activities have become more abundant. At the same time, the convenience of network communication makes it easier for us to understand the development status of all parts of the world, which brings the geographical diversity of our outdoor activities. Among the many types of points of interest, we may not be able to choose the next point of interest we want to go to for a while, and the point of interest recommendation algorithm helps us choose the right point of interest. The traditional recommendation model based on the collaborative filtering method uses items that other users have browsed in the past to recommend for the current user, and uses the items that the current u...

AI Technical Summary

Problems solved by technology

The current point-of-interest recommendation method is mainly to integrate the context information of the point of interest into the collaborative filtering algorithm and the model-based algorithm. Extracting and expressing behaviors seriously reduces the accuracy and personalization of point-of-interest recommendations

Images