複数センサデータの自動可視化に向けたユーザ操作機械学習の予備評価

Keita Fujino, Moe Matsuki, Sozo Inoue, Tom Shibata,
日本知能情報ファジィ学会九州支部学術講演会予稿集
(Not Available)
(Not Available)
29-32
2016-12-10
Kumamoto
http://soft-kyushu.org/2016/conf/index.html
The purpose of this study is to automatically visualize multiple sensor data according to the user. Therefore, the user operated to perform zooming and sensor selection on visualized data, and we analyze where the user is examining on. The operation was estimated by using random forest with features from the data as the explanatory variables and the user operations as the objective variable, and was analyzed using important factors. As a result, it was found that features such as “Number of Samples” of data and “Maximum Value” and “Variance” on the right side of visualized data were high importance.

Data Files