INTERNATIONAL JOURNA L OF ENGINEERING SCI ENCES & RESEARCH TECHNOLOGY Error -Correcting Output Code & Pessimistic Decoding Technique
Devangini Dave, M. Samvatsar, Praveen Kumar Bhanodia · 2012
Error correcting output codes (ECOC) represent a su ccessful extension of binary classifiers to address the multiclass problem. Lately, the ECOC framework was extended from the binary to the tern ary case to allow classes to be ignored by a certain classifier, allowing in this w ay to increase the number of possible dichotomies t o be selected. we show that by a special treatment procedure of ze ros, and adjusting the weights at the rest of coded posi accuracy of the system can be increased. Besides, w e extend the main state -of-art decoding strategies from the binary to the ternary case, and use two novel appro aches: Laplacian and Pessimistic Beta Density Proba bility arch is to show that the ternary decoding technique s proposed outperform the standard In this survey paper, we present an open source Err or -Correcting Output Codes (ECOC) library. The ECOC framework is a powerful tool to deal with multiclass categorization problems. This library co ntains both one, one -versus-all, dense random, sparse random, DECOC, forest ONE) and decoding designs (hamming, Euclidean, inverse hamming, laplacian, � -density, attenuated, loss based, and loss -weighted) with the parameters defined by the authors, as well as the option to include your own coding, decoding, and ba se classifier.