A Method of Activity Classification based on Behavioral Similarity and its Accuracy Verification
-
- Chikaraishi Makoto
- Hiroshima University
-
- Zhang Junyi
- Hiroshima University
-
- Fujiwara Akimasa
- Hiroshima University
Bibliographic Information
- Other Title
-
- 行動の類似性に基づく活動分類手法の提案及び精度検証
- 行動の類似性に基づく活動分類手法の提案及び精度検証--活動開始時刻を例に
- コウドウ ノ ルイジセイ ニ モトズク カツドウ ブンルイ シュホウ ノ テイアン オヨビ セイド ケンショウ カツドウ カイシ ジコク オ レイ ニ
- 活動開始時刻を例に
- A Case Study of Activity Start Time
Search this article
Description
This paper proposes a method for activity classification based on behavioral similarity which is defined by the influencing structure of observed variables on a behavioral aspect of interest. The method takes into account the fact that individual behavior changes from day to day, i.e., an observed behavior at a certain moment may not represent typical behavior. An empirical analysis was conducted focusing on activity start time. In order to check the predictive ability of proposed activity classification method, its accuracy was compared with a priori activity classification based on the composition of sum of squares. The results indicate that the prediction accuracy of proposed method is much higher than that of a priori classification.
Journal
-
- Journal of the City Planning Institute of Japan
-
Journal of the City Planning Institute of Japan 44.3 (0), 481-486, 2009
The City Planning Institute of Japan
- Tweet
Keywords
Details 詳細情報について
-
- CRID
- 1390001204214266496
-
- NII Article ID
- 10025844664
- 130006949215
-
- NII Book ID
- AN10276504
-
- ISSN
- 21850593
- 09160647
- 1348284X
-
- NDL BIB ID
- 10404980
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- NDL Search
- Crossref
- CiNii Articles
- KAKEN
- OpenAIRE
-
- Abstract License Flag
- Disallowed