An online taxi recommendation method based on collaborative filtering

Abstract

The invention discloses an online taxi calling recommending method based on collaborative filtering; the method comprises the following steps: 1, original data preprocessing, real GPS data needs a preprocessing step mainly comprising passenger-carrying origin and terminal point extraction; 2, a gridding step comprising area filtering, gridding segmentation and user filtering; 3, recommending algorithm, introducing a collaborative filtering algorithm, using a taxi-based algorithm and a latent factor model to provide a recommending result, and forming a reasonable score matrix; 4, evaluation and result application: using conventional recommendation topN recommend to examine the result by using accuracy and recall rate as indexes. The novel method aims to parse GPS data of a taxi in real time, and parse the running state so as to form and carve the score matrix of taxi behaviors, thus providing customized no-load passenger searching location recommendations for the taxi. The advantages are that the conventional collaborative filtering method is introduced, so the recommendations are customized; in addition, the model based on mobile characteristics can comply with reality in a more accurate manner.

Description

technical field [0001] The invention belongs to the field of computer applications, in particular to a technique for recommending places for taxi drivers to find customers. [0002] technical background [0003] With the development of the city, the traffic problem bears the brunt, and the difficulty of taking a taxi and finding passengers is serious. Taxis are generally equipped with GPS devices, and traffic terminals can obtain taxi location information and passenger status information. It is of great practical significance to provide unloaded taxis with places to find passengers through the traffic recommendation system, which can not only effectively reduce the number of taxis Empty time increases the operating efficiency of taxis, thereby increasing their income. [0004] The problem of location recommendation for taxis can be summarized as follows: Given a time point and location information, several location recommendations are given for specific taxis. In recent yea...

AI Technical Summary

Problems solved by technology

The disadvantages of this most popular recommendation are: 1. The most popular location recommendation lacks the evaluation of competing operations and is not accurate enough; 2. For any taxi, the recommendation results are the same and lack of personalization

Images