user single interest point identification method based on space-time data

The invention discloses a spatio-temporal data-based user single interest point identification method. The method comprises the following steps: 1) performing error data filtering; 2) carrying out DBSCAN clustering; performing clustering judgment; 3) calculating related variables; and 4) adopting a family or work address single-class algorithm. Home working addresses of most users can be mined only through the journey data. According to the method, the main interest points of the user are mined as potential family working addresses through DBSCAN clustering of the main interest points. Furthermore, all users who do not meet the clustering condition are filtered, and the flow and the calculation cost are reduced. Furthermore, related variables are accurately calculated to be used for subsequent algorithm judgment, and the judgment precision is improved. Furthermore, the probability calculation is carried out on the single cluster to determine whether the cluster is a family, a working address or not. 本发明公开了一种基