PROGRAM FOR ESTIMATING BOARDING INFORMATION ON USER FROM COMMUNICATION LOG OF PORTABLE TERMINAL, DEVICE, AND METHOD

• Main Authors: KOBAYASHI AREI, KUROKAWA MORI
• Format: Patent
• Language: English; Japanese
• Published: 24.10.2019
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC COMMUNICATION TECHNIQUE; ELECTRIC DIGITAL DATA PROCESSING; ELECTRICITY; PHYSICS; TELEPHONIC COMMUNICATION

To provide a program and the like for constructing a learning model of a machine learning engine that estimates boarding information of a boarding machine of a user by using a communication log of a portable terminal held by the user.SOLUTION: Functioned as a departure time table in which a departure time of each boarding machine and boarding information are defined in each point, communication log accumulation means that accumulates a communication log containing a time and a position of each portable terminal in each point,and trip log extraction means of extracting, by using the communication log accumulation means, a pair of a preceding communication log in which a distance between positions of the communication log which is continued in a time sequence is a prescribed length or more and a post communication log as a trip log, a machine learning engine learns in associating "the preceding communication log" as a departure point as teacher data with "the boarding information" in which a time of the preceding communication log is contained in a prescribed preceding time range of the departure time of the departure time table.SELECTED DRAWING: Figure 2 【課題】ユーザによって所持される携帯端末の通信ログを用いて、当該ユーザの搭乗機の搭乗情報を推定する機械学習エンジンの学習モデルを構築するプログラム等を提供する。【解決手段】拠点毎に、各搭乗機の出発時刻及び搭乗情報が規定された出発時刻表と、拠点毎に、各携帯端末の時刻及び位置を含む通信ログを蓄積した通信ログ蓄積手段と、通信ログ蓄積手段を用いて、時系列に連続する通信ログの位置間の距離が所定長以上となる先方通信ログ及び後方通信ログの組を、トリップログとして抽出するトリップログ抽出手段として機能させ、機械学習エンジンは、教師データとして、出発拠点となる「先方通信ログ」と、当該先方通信ログの時刻が、出発時刻表の出発時刻の所定前時間範囲に含まれる「搭乗情報」とを対応付けて学習する。【選択図】図２