发明名称 DEVICE AND METHOD FOR DESTINATION PREDICTION
摘要 A device for constructing a hierarchical destination prediction mode includes an extracting unit configured to receive a user's historical location data and extract input data and output data from the user's historical location data. The input data includes a leaving location and associated time for the user and the output data includes an arriving location for the user. A determining unit is configured to determine the user's possible activities of each location and their original probabilities based on the input data and the output data, and to adjust the original activity probabilities based on the user's arriving patterns and leaving patterns. A prediction model obtaining unit is configured to obtain the hierarchical destination prediction model based on the input data and the adjusted probabilities. A high layer of the hierarchical destination prediction model represents activities and their transitions, and a low layer of the model represents locations of the activities.
申请公布号 US2014207725(A1) 申请公布日期 2014.07.24
申请号 US201313938630 申请日期 2013.07.10
申请人 NEC (China) Co., Ltd. 发明人 LI MAN;HU WEISONG;LIU XIAOWEI
分类号 G06N5/04 主分类号 G06N5/04
代理机构 代理人
主权项 1. A device for constructing a hierarchical destination prediction model, the device, including a processor and memory, comprising: an extracting unit configured to receive user's historical location data and extract input data and output data from the user's historical location data, the input data including a leaving location and associated time for the user, the output data including an arriving location for the user; a determining unit configured to determine the user's possible activities of each location and their original probabilities based on the input data and the output data, and to adjust the original activity probabilities based on a user's arriving patterns and leaving patterns; and a prediction model obtaining unit configured to obtain the hierarchical destination prediction model based on the input data and the adjusted probabilities, wherein a high layer of the hierarchical destination prediction model represents activities and their transitions and a low layer of the hierarchical destination prediction model represents locations of the activities.
地址 Beijing CN