Abstract
Data heterogeneity creates the complexity for understanding the different domains. Most of domains require sophisticated algorithm to understand the complexity. To handle the data heterogeneity problem, several methods have been proposed (e.g., Support Vector Machine (SVM), Decision Tree, Random Forest, Liner regression etc.). However, these existing methods suffer due to accuracy, outlying. In this paper, we introduce Pair Data Division (PDD) with support of SMV for handling the data heterogeneity. PDD divides the data into the pairs then further pairs are processed using SVM. The PDD algorithm is implanted by using Python and PyCharm is used as Interface Development Environment (IDE). Furthermore, training datasets will be given to PDD for testing purpose. Based on the testing results, we confirm that our proposed algorithm provides better accuracy as compared to known algorithms: Random Forest and Linear regression when handling the data heterogeneity.