An Efficient Ensemble Based Model for Data Classification

Classification techniques are very useful in prediction, decision making of data. Classification technique is being used in IT industry, government organization for classifying the data and to infer the knowledge for decision making or knowledge discovery. There are various classification algorithms like Logistic regression, Naïve Bayes, Decision tree, Support vector machine, and K-Nearest Neighbors etc. these algorithms have limitations like high time requirement to train the data and data sensitivity, performance. In this paper we propose a hybrid model i.e. ensemble and stacking technique to overcome the limitations of existing classification techniques. Hybridization enhances the accuracy and precision value and also over comes the limitations of existing classification algorithms. Most of the times when classifying textual data using individual learners such as Naive Bayes, Support vector machine, Logistic Regression they have a high bias or high variance because of which the performance is reduced. Ensemble methods can be used to increase the performance by combining several weak learners together in order to create a strong learner. The main aim is to implement a hybrid ensemble method for classification on the given dataset and perform comparative study analysis with some of the existing algorithms by considering performance indicators such as precision, recall, F1-score and support.