| Item type |
SIG Technical Reports(1) |
| 公開日 |
2018-11-08 |
| タイトル |
|
|
タイトル |
Extracting Class-specific Sequential Pattern for Continuous Glucose Monitoring |
| タイトル |
|
|
言語 |
en |
|
タイトル |
Extracting Class-specific Sequential Pattern for Continuous Glucose Monitoring |
| 言語 |
|
|
言語 |
eng |
| キーワード |
|
|
主題Scheme |
Other |
|
主題 |
MBL |
| 資源タイプ |
|
|
資源タイプ識別子 |
http://purl.org/coar/resource_type/c_18gh |
|
資源タイプ |
technical report |
| 著者所属 |
|
|
|
IBM Research - Tokyo |
| 著者所属 |
|
|
|
IBM Research - Tokyo |
| 著者所属 |
|
|
|
Fujita Health University |
| 著者所属 |
|
|
|
The Dai-ichi Life Insurance Company, Limited |
| 著者所属 |
|
|
|
Fujita Health University |
| 著者所属(英) |
|
|
|
en |
|
|
IBM Research - Tokyo |
| 著者所属(英) |
|
|
|
en |
|
|
IBM Research - Tokyo |
| 著者所属(英) |
|
|
|
en |
|
|
Fujita Health University |
| 著者所属(英) |
|
|
|
en |
|
|
The Dai-ichi Life Insurance Company, Limited |
| 著者所属(英) |
|
|
|
en |
|
|
Fujita Health University |
| 著者名 |
Masaki, Ono
Takayuki, Katsuki
Masaki, Makino
Kyoichi, Haida
Atsushi, Suzuki
|
| 著者名(英) |
Masaki, Ono
Takayuki, Katsuki
Masaki, Makino
Kyoichi, Haida
Atsushi, Suzuki
|
| 論文抄録 |
|
|
内容記述タイプ |
Other |
|
内容記述 |
Continuous glucose monitoring (CGM) is temporal time-series data that has been available for approximately 10 years thanks to the invention of a device with low measurement error. Understanding the time series variation of glucose helps you treat specific patient groups better by understanding their lifestyles. Therefore, we propose a method of extracting characteristic sequential patterns from given pairs of labels and sequences. First, we apply time-series clustering to transform CGM value sequences into cluster id sequences. Next, we apply sequential pattern mining to extract frequently occurred sequences. Finally, we evaluate each frequent sequence based on their correlation to a specific class. We experimented with two datasets that is manually created and one real CGM dataset to prove our method is effective. |
| 論文抄録(英) |
|
|
内容記述タイプ |
Other |
|
内容記述 |
Continuous glucose monitoring (CGM) is temporal time-series data that has been available for approximately 10 years thanks to the invention of a device with low measurement error. Understanding the time series variation of glucose helps you treat specific patient groups better by understanding their lifestyles. Therefore, we propose a method of extracting characteristic sequential patterns from given pairs of labels and sequences. First, we apply time-series clustering to transform CGM value sequences into cluster id sequences. Next, we apply sequential pattern mining to extract frequently occurred sequences. Finally, we evaluate each frequent sequence based on their correlation to a specific class. We experimented with two datasets that is manually created and one real CGM dataset to prove our method is effective. |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AA11515904 |
| 書誌情報 |
研究報告高度交通システムとスマートコミュニティ（ITS）
巻 2018-ITS-75,
号 18,
p. 1-8,
発行日 2018-11-08
|
| ISSN |
|
|
収録物識別子タイプ |
ISSN |
|
収録物識別子 |
2188-8965 |
| Notice |
|
|
|
SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc. |
| 出版者 |
|
|
言語 |
ja |
|
出版者 |
情報処理学会 |
IPSJ-ITS18075018.pdf (1.1 MB)
