Building a Machine Learning Model for Unstructured Text Classification: Towards Hybrid Approach

Abstract

Text analytics is one of the major research domains. Many researchers are working on business intelligence to create decision-making systems of high accuracy. We are enjoying the result from these machine learning models but these systems are hungry for data. Too much data is available to deal with as it is arriving at incredible speed. Problem is that 90% of data is unstructured, and it is very difficult to tackle this raw data. Data scientists are trained to deal with categorical and numerical data only. The idea is to create a machine learning model for unstructured text categorization. The proposed model is going to use the k-means algorithm for text clustering followed by a deep neural network for classification. This hybrid clustering and classification (HCC) model is a combination of state of the art algorithm k-means and very hot concept Deep Learning. The main focus of this research is the unstructured text that can be utilized for various natural language processing (NLP) applications.