EXISTENCE ANALYSIS METHOD, SYSTEM, AND PROGRAM

• Main Authors: UCHIYAMA MASASHI, TAKAYA NORIKO, NAGANO SHOICHI, ICHIKAWA YUSUKE
• Format: Patent
• Language: English; Japanese
• Published: 19.01.2015
• Subjects: CALCULATING; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

PROBLEM TO BE SOLVED: To highly precisely estimate a distribution of departure probabilities.SOLUTION: A data acquisition unit 24 acquires data items of plural customers. A parameter initial value designation unit 26 designates an initial value of a distribution parameter, which represents a distribution of departure probabilities of a phenomenon and relates to a steady factor x, for each of steady factors x, and designates an initial value of a distribution parameter, which represents a distribution of departure probabilities of a phenomenon and relates to a non-steady factor y, for each of non-steady factors y. A model estimation unit 30 determines a factor, to which a customer belongs, for each of customers, updates the distribution parameters of the steady factors x, updates the distribution parameters of the non-steady factors y, determines the distribution parameter of each of the steady factors x on the basis of the repeatedly updated distribution parameters of the steady factors x, and estimates a parameter that determines the distribution parameters of the non-steady factors y on the basis of the distribution parameters of the non-steady factors y repeatedly updated by non-steady factor distribution parameter updating means. 【課題】精度良く離脱確率の分布を推定することができる。【解決手段】データ取得部２４により、複数の顧客のデータを取得し、パラメータ初期値設定部２６により、定常要因ｘの各々について、定常要因ｘによる現象の離脱確率の分布を表す分布パラメータの初期値を設定すると共に、非定常要因ｙの各々について、非定常要因ｙによる現象の離脱確率の分布を表す分布パラメータの初期値を設定し、モデル推定部３０により、顧客の各々について所属する要因を決定し、定常要因ｘの分布パラメータを更新し、非定常要因ｙの分布パラメータを更新し、繰り返し更新された定常要因ｘ毎の分布パラメータに基づいて、定常要因ｘ毎の分布パラメータを決定し、非定常要因分布パラメータ更新手段により繰り返し更新された非定常要因ｙ毎の分布パラメータに基づいて、非定常要因ｙ毎の分布パラメータを決定するパラメータを推定する。【選択図】図１