Knowledge Discovery from Huge Data Ensemble by an Integration of Automatic Data Selection and Pattern Extraction
| Project/Area Number |
25280085
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Partial Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Kyushu University |
|---|
Principal Investigator |
Suzuki Einoshin 九州大学, システム情報科学研究科(研究院, 教授 (10251638)
|
|---|
| Project Period (FY) |
2013-04-01 – 2016-03-31
|
| Project Status |
Completed (Fiscal Year 2015)
|
| Budget Amount *help |
¥17,160,000 (Direct Cost: ¥13,200,000、Indirect Cost: ¥3,960,000)
Fiscal Year 2015: ¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2014: ¥5,460,000 (Direct Cost: ¥4,200,000、Indirect Cost: ¥1,260,000)
Fiscal Year 2013: ¥7,540,000 (Direct Cost: ¥5,800,000、Indirect Cost: ¥1,740,000)
|
|---|
| Keywords | 巨大データ集合 / 自動データ選択 / パターン抽出 / データマイニング / 巨大データ集合体 |
|---|
| Outline of Final Research Achievements |
To discover knowledge from a large number of huge data sets, we invented, developed, and implemented 4 highly novel methods that integrate automatic data selection and pattern discovery. The method that discovers cluster distribution meta-patterns exhibited high recalls and precisions under difficult conditions of high noise contamination and ambiguous and mutually overlapping cluster boundaries and was proved to be time-efficient. The method that discovers directional non-zero weight meta-patterns, which is based on multi-task classification based on sparse modeling, showed its practicability on various kinds of data including facial expression data measured with Kinect. The method that hierarchically clusters linear classifiers and the method that evaluates and discovers general classification rules each holding true in its respective data set showed their effectiveness on various synthetic and real data.
|
Report
(4 results)
Research Products
(38 results)
