IMPROVING PREDICTIONS USING QUALITATIVE PARAMETERS

Abstract

 Selecting appropriate parameters while making any prediction model is a tedious task. Often, while constructing a prediction model, categorical variables are ignored. If we include more qualitative parameters for prediction, the observed results will have more accuracy. Neural networks help in a proper learning methodology which utilizes the concept of machine learning. When prediction is to be made, the human behavioral patterns hamper the test results as it plays a crucial role in any decision making. Employing qualitative parameters in decision making, accurate conjectures are possible. Qualitative parameters are considered fuzzy in nature and neural networks, which is one of the major components of the soft computing, works very well with incomplete data. In this paper we have discussed how qualitative parameters will help in improving, the prediction accuracy, and the decision making logic, to make predictive models more sustainable and robust.