Presentation | 2011-01-25 Learning from Partially Supervised Data Stream Using Past Instances To Label Clusters(Knowledge-Based Software Engineering) Tatsuya KOSAKA, Yoshiaki YASUMURA, Kuniaki UEHARA, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | This report presents a learning method for a partially supervised data stream that contains few labeled positive and unlabeled data. Instances in a data stream are divided into chunks, which contains a fixed size of instances. This method is based on the ensemble approach that builds classifiers from chunks. Then, in order to learn from a chunk, this method extracts reliable negative instances from unlabeled data using k-means clustering with labeled positives in the past chunks. To detect concept change in a data stream, clusters are assigned the weights based on the distance between the clusters containing labeled positives in the coming chunk and the clusters containing it in the past ones. The experiment results using an artificial data stream show the validity of the proposed method. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | data stream / partially supervised learning / concept change / K-means clustering / past instances |
Paper # | KBSE2010-41 |
Date of Issue |
Conference Information | |
Committee | KBSE |
---|---|
Conference Date | 2011/1/17(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Knowledge-Based Software Engineering (KBSE) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Learning from Partially Supervised Data Stream Using Past Instances To Label Clusters(Knowledge-Based Software Engineering) |
Sub Title (in English) | |
Keyword(1) | data stream |
Keyword(2) | partially supervised learning |
Keyword(3) | concept change |
Keyword(4) | K-means clustering |
Keyword(5) | past instances |
1st Author's Name | Tatsuya KOSAKA |
1st Author's Affiliation | Department of Computer Science and Systems Engineering, Faculty of Engineering, Kobe University() |
2nd Author's Name | Yoshiaki YASUMURA |
2nd Author's Affiliation | College of Engineering, Shibaura Institute of Technology |
3rd Author's Name | Kuniaki UEHARA |
3rd Author's Affiliation | Department of Computer Science and Systems Engineering, Faculty of Engineering, Kobe University |
Date | 2011-01-25 |
Paper # | KBSE2010-41 |
Volume (vol) | vol.110 |
Number (no) | 386 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |