Development of a feature construction method based on sequential pattern evaluation index and temporal pattern extraction
| Project/Area Number |
24500175
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Bunkyo University |
|---|
Principal Investigator |
ABE Hidenao 文教大学, 情報学部, 講師 (00397853)
|
|---|
| Project Period (FY) |
2012-04-01 – 2015-03-31
|
| Project Status |
Completed (Fiscal Year 2014)
|
| Budget Amount *help |
¥4,030,000 (Direct Cost: ¥3,100,000、Indirect Cost: ¥930,000)
Fiscal Year 2014: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2013: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2012: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
|
|---|
| Keywords | 系列パターン評価指標 / 時系列クラスタリング / 属性構築 / 系列パターンマイニング / 分類学習 / 転移学習 / 系列データマイニング |
|---|
| Outline of Final Research Achievements |
This study aims to develop a feature construction method from complex dataset, which consist of not only numerical and nominal values but also their temporal values. However, conventional feature extraction methods have been only developed for each kind of values for constructing features for performing some statistical methods and machine learning methods.
In this study, existing importance evaluating indices from the natural language processing field and sequential pattern evaluation indices were formulated to an integrated viewpoint for evaluating sequential patterns. Then, by combining these indices with temporal pattern extraction method, the method has enable to extract temporal causalities for understanding user behaviors on Web clickstream and temporal text corpus on SNS from more data-centric viewpoint.
|
Report
(4 results)
Research Products
(12 results)
