Development and Application of an Imbalanced Data Classifier
Project/Area Number |
15K00323
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Research Field |
Intelligent informatics
|
Research Institution | Doshisha University |
Principal Investigator |
Ohsaki Miho 同志社大学, 理工学部, 教授 (30313927)
|
Project Period (FY) |
2015-04-01 – 2018-03-31
|
Project Status |
Completed (Fiscal Year 2017)
|
Budget Amount *help |
¥3,640,000 (Direct Cost: ¥2,800,000、Indirect Cost: ¥840,000)
Fiscal Year 2017: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2016: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2015: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
|
Keywords | 不均衡データ分類 / 混同行列 / カーネルロジスティック回帰 / 最小分類誤り学習 / 一般化確率的勾配法 |
Outline of Final Research Achievements |
In a wide range of domains such as cancer diagnosis, vehicle accident prediction, etc., there is a high demand for the classification of a small number of emergent instances (minority class) and a large number of ordinary instances (majority class). However, the imbalance of the two classes causes overlooking minorities. Conventional solutions for this were domain-specific and difficult to control the balance of performance between the classes. We therefore aim at the development of an imbalanced data classifier which is of high versatility and achieves the balance control and the improvement of performances. The proposed method is based on kernel logistic regression, minimum classification error and generalized probabilistic descent, and confusion matrix. The superiority of the proposed method to the conventional ones was confirmed by the evaluation experiments. We finally published an academic journal paper to report all this research results.
|
Report
(4 results)
Research Products
(10 results)