Investigation and Development of Methods to Solve Multi-Class Classification Problems

摘要: Most of the classification problems frequently encounter a multi class predicament and offers good scope for research. This paper has comprehensive approach to available multi-class technique using Artificial Neural Networks then introduces new algorithm overcome demerits former. In addition, combining ANN chameleon clustering is suggested validated. An SVM model above also proposed sufficiently tested with typical example i.e. Image Segmentation. Also, permutation effects prevailing in Half –against-Half efficiently tackled by developing an “circular shift strategy” employing same. The use methods improve its efficiency discussed. All mentioned models are extensively analyzed results presented. It found that method effective alternative existing consistent performance.