GPSログデータを用いた訪日外国人旅行者の訪問パターンの分析手法の開発  [in Japanese] Development of the Combination Analysis Method of Visited Places of Foreign Visitors to Japan by GPS Log Data  [in Japanese]

Access this Article

Search this Article

Author(s)

Abstract

<p>本研究は,訪日外国人旅行者の訪問パターンの特性把握をGPSログデータを用いて行うことを目的とする.論文では,はじめに訪問地の組み合わせである訪問パターンの抽出にトピックモデルを用いることを説明するとともに,潜在クラス分析との差異を明らかにした.トピックモデルは教師データなしの機械学習(セグメンテーション手法)の1つと位置づけられ,セグメントの確率的導出過程が明示でき,セグメント相互が排他的でない特色を有する.分析では,訪日外国人向けナビゲーションアプリにより取得できたGPSログ情報を用いて,トピックモデルによって訪問場所の組み合わせパターンを幾つかのクラスに分類した.</p>

<p>The purpose of this paper was to develop the combination analysis method of visited places by foreign visitors to Japan. For identification of combination patterns, the topic model was applied using GPS tracking data of smartphone applications of foreign tourists. This model was supposed as a method of identifying segments of destinations and visitors simultaneously. The number of unique users collected in 2015 was 55,199. We had data cleaning for bias caused due to sampling ratio and the noise of GPS positioning data. In the estimation, it was found appropriate to set the number of topics as six. The visited rates and visitors' topic distribution of each zone were also estimated by this model.</p>

Journal

  • Transport Policy Studies' Review

    Transport Policy Studies' Review 20(0), 020-029, 2018

    Japan Transport Research Institute

Codes

  • NII Article ID (NAID)
    130006725349
  • NII NACSIS-CAT ID (NCID)
    AA11287548
  • Text Lang
    JPN
  • ISSN
    1344-3348
  • NDL Article ID
    029123171
  • NDL Call No.
    Z71-C180
  • Data Source
    NDL  J-STAGE 
Page Top