論 文Papers

CONFERENCE (DOMESTIC)

活動量計テータを用いた同行検出手法の研究

川尻 亮真(東京大学), 坪内 孝太, 荒木 青伊良(東京大学), 下坂 正倫(東京大学)

2014年度人工知能学会全国大会(第28回), 2014/5

Category:

データサイエンス (Data Science)

Abstract:
This paper proposes an innovative way to detect working relationships by using only the step tracking data acquired from pedometers like Fitbit. The idea makes the cost of working-relationship detection much lower than that of previous approaches. We can find out if people have a working relationship and spend their daily lives together by making them wear a pedometer. Results of an experiment in Japan showed that this approach is very effective and practical. An organizations profile can be written automatically by analyzing the data in correctly.
Download:

活動量計テータを用いた同行検出手法の研究(PDF)