PREDICTED MODEL LEARNING SYSTEM, EVENT PREDICTING SYSTEM, METHOD FOR LEARNING PREDICTED MODEL, AND PROGRAM,申请号JP20100057892-传众专利搜索

PREDICTED MODEL LEARNING SYSTEM, EVENT PREDICTING SYSTEM, METHOD FOR LEARNING PREDICTED MODEL, AND PROGRAM

申请公布号：JP2011192040(A)

申请号：JP20100057892

申请日期：2010.03.15

申请公布日期：2011.09.29

申请人：

KDDI CORP

发明人：KUROKAWA MORI;KAMISAKA DAISUKE

分类号：G06N5/04;G06Q10/04

主分类号：G06N5/04

摘要：PROBLEM TO BE SOLVED: To achieve precise event occurrence prediction, by providing a technique for dividing time space easily and appropriately. SOLUTION: A mobile station terminal 1 includes an event-related information storage unit 100 for correlating event data of a prescribed event, position data at the occurrence of a prescribed event, and time data at the time of occurrence of a prescribed event to event-related information for storage; a time space predicted model information storage unit 110 for storing time space predicted model information, relating to a prescribed event for each time space region assumed from a time axis and a space axis; a region division necessity determining unit 140 for determining whether new generation of a time space region is required, by referring to the time space predicted model information, when acquiring the event-related information; and a time space predicted model information updating unit 150 for updating the time space predicted model information, when the region division necessity determining unit 140 determines that new generation of a time space region is required. COPYRIGHT: (C)2011,JPO&INPIT