Information intelligent recommendation method, system and storage medium

Abstract

The present invention provides an intelligent recommendation method for information, data set construction: constructing user training data set U D1 and the test dataset U D2 , construct the training data set U of the gas station G1 and the test dataset U G2 ; Preprocess the training data set and the test data set respectively, train the neural network model to be trained, and obtain the trained model M 21 , M 22 and M 23 ; Obtain the recommended data of each gas station, sort the gas stations based on the recommended data of each gas station, and obtain the final gas station recommendation information. The present invention improves the training speed and accuracy of the model by improving the activation function of the neural network model, and realizes accurate recommendation of gas stations for users by comprehensively considering multi-dimensional information such as actual factors, user consumption, user historical behavior statistics, and user information. , providing accurate user maintenance for gas stations, effectively improving the user experience and payment rate of gas stations.

Description

technical field [0001] The invention relates to information recommendation, in particular to intelligent recommendation of gas station related information. Background technique [0002] With the popularization of automobiles, more and more users choose to travel by self-driving, so the need to refuel may appear at any time. How to find a suitable gas station for users is an urgent problem that needs to be solved. When recommending gas stations in the existing technology, usually only single-dimensional factors such as distance are considered, and there are problems such as fuzzy recommendation rules for gas stations and inaccurate recommendation results, that is, the current gas station recommendation method does not comprehensively consider the user's historical usage and the possible benefits and benefits of gas stations. The prior art still has problems such as chaotic recommendation order and failure to achieve the purpose of recommendation well. [0003] At the same t...

AI Technical Summary

Problems solved by technology

When recommending gas stations in the existing technology, usually only single-dimensional factors such as distance are considered, and there are problems such as fuzzy recommendation rules for gas stations and inaccurate recommendation results, that is, the current gas station recommendation method does not comprehensively consider the user's historical usage Dimensions such as benefits and discounts that may exist at gas stations

There are still problems in the existing technology that the order of recommendation is chaotic, and the purpose of recommendation cannot be well achieved.

[0003] At the same time, the existing technology does not have a complete rule system that can be provided to gas stations to help gas stations maintain high-quality users and increase the payment rate of gas stations

Images