Name
Abstract | ||
|---|---|---|
This paper presents a new study on a method of designing a multi-class classifier: Data-driven Error Correcting Output Coding (DECOC). DECOC is based on the principle of Error Correcting Output Coding (ECOC), which uses a code matrix to decompose a multi-class problem into multiple binary problems. ECOC for multi-class classification hinges on the design of the code matrix. We propose to explore the distribution of data classes and optimize both the composition and the number of base learners to design an effective and compact code matrix. Two real world applications are studied: (1) the holistic recognition (i.e., recognition without segmentation) of touching handwritten numeral pairs and (2) the classification of cancer tissue types based on microarray gene expression data. The results show that the proposed DECOC is able to deliver competitive accuracy compared with other ECOC methods, using parsimonious base learners than the pairwise coupling (one-vs-one) decomposition scheme. With a rejection scheme defined by a simple robustness measure, high reliabilities of around 98% are achieved in both applications. |
Year | DOI | Venue |
|---|---|---|
2008 | 10.1016/j.patcog.2007.05.020 | Pattern Recognition |
Keywords | Field | DocType |
support vector machine,gene expression classification,handwritten numeral recognition,code matrix,multi-class classification,ecoc method,compact code matrix,multi-class classification hinge,data-driven error correcting output,base learner,proposed decoc,error correcting output coding ecoc,data-driven error correcting output coding decoc,multi-class classifier,error correcting output coding,multi-class problem,data-driven decomposition,multi class classification,gene expression | Pairwise comparison,Pattern recognition,Computer science,Support vector machine,Error detection and correction,Robustness (computer science),Coding (social sciences),Artificial intelligence,Classifier (linguistics),Code (cryptography),Machine learning,Multiclass classification | Journal |
Volume | Issue | ISSN |
41 | 1 | Pattern Recognition |
Citations | PageRank | References |
38 | 1.39 | 21 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Jie Zhou | 1 | 139 | 10.45 |
| Hanchuan Peng | 2 | 3930 | 182.27 |
| Ching Y. Suen | 3 | 7569 | 1127.54 |