Open AccessA Prediction Model of Online Car-Hailing Demand Based on K-means and SVR
Author(s) -
Bang Chen,
Shenghan Zhou,
Houxiang Liu,
Xinpeng Ji,
Yue Zhang,
Wenbing Chang,
Yiyong Xiao,
Xing Pan
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1670/1/012034
Subject(s) - division (mathematics) , cluster analysis , support vector machine , computer science , set (abstract data type) , data set , data mining , research object , operations research , artificial intelligence , geography , mathematics , arithmetic , regional science , programming language
The paper proposed a prediction model of online car-hailing demand based on K-means and support vector regression (SVR) methods. In the past few years, online car-hailing market demand has grown rapidly, and prediction of rapid demand growth has become a hot topic. This study takes the initial longitude and latitude of online car-hailing orders as the eigenvalues for K-means clustering. The clustering results are taken as the result of area division. The number and size of potential demand areas could be determined automatically. This method of area division solves the shortcomings of traditional artificial meshing division and existing administrative division methods. The model takes a small-sample data set as the application object and uses the SVR method for data regression. Finally, we conduct an empirical study on Didi’s real data set in the core area of Chengdu City, China. The final experimental results suggest that the area division method based on K-means is reasonable and that the demand prediction model based on K-means and SVR is effective.