Improved Naive Bayes Classification for Joint Investment Plan

Abstract

Large scale investments are mostly done by joint investors in different countries. Most of these investments involve collaboration with financial institutes of different countries. As the aspiration of governments to development their countries, they encourage investments. Financial institutes, at the same time, will set a guideline to decide with whom they will share the investment and collaborate based on profit maximization target. In this paper we are considering individual investors to collaborate with the financial institutes. Naïve Bayes is an ideal approach to aid the approval or rejection of this collaboration by the decision maker. The approach assumes independencies among the variables. However, this assumption may not always be realistic. Hence, this paper uses a method to improve the accuracy of Naïve Bayes approach by using a learning structure of feature variables in the model and apply it to joint investment plan applications. The introduction and use of new applied problem is not only helpful to show the application of the field but also attract researchers from social science to apply and use Bayes based methods which in turn contribute the development of the field with new insights.