Adaptive rule based multi class svm classification of big biological data

Biological field of research have as of now delivered huge measures of important biological information that is trying to examine because of its high dimensionality and complexity. With this huge amount of data there are several problems during classification of data such as: noisy instances, over fitting as well as class imbalance. This will influence accuracy as well as the efficiency of supervised learning techniques. This paper proposes a data-adaptive rule-based classification framework for biological big data classification which produces generates relevant rules by finding adaptive partitions. Rule based classifier is the proposed system, which is consolidation of random subspace as well as boosting approaches. Classification rules development that does not have global optimization, system utilizes J48 Decision Tree and KNN algorithm. Random subspace is used to avoid over fitting problem, boosting approach is used for solving problem of noisy instances classification and finally J48 decision tree is deal with class imbalance problem. With J48 decision tree, evaluation of rules is done from training data as well as with KNN, misclassified instances are analyzed. The developed method will be contrasted with various rule-based and machine learning classifiers named as Support Vector Machine (SVM).