GPS (global positioning system) geographic position data mining based user classification method

The invention relates to a GPS (global positioning system) geographic position data mining based user classification method which comprises the following steps of: preprocessing an initial data set, extracting data attribute required by an algorithm and summing up and sorting out the data attribute by taking users as units; mining staying regions of the users through a FindStayArea algorithm, and finding out large-range activity regions (such as a workplace region, a residential quarter and the like) of the users; then mining staying positions, namely activity positions (such as a building in the workplace region, a stadium in the residential quarter and the like), of the users through a FindStayPoint algorithm; and finally, classifying the users by combining the staying positions of the users with a UserClassify algorithm. According to the method, original GPS geographic position data of the users is subjected to hierarchical processing, so that the staying positions in which the users often perform activities can be mined, the users also can be classified by utilizing a mining result, a user group with highest similarity is found, and the value of the GPS geographic position data is fully mined.