Improved Text Analysis Approach for Predicting Effects of Nutrient on Human Health using Machine Learning Techniques

Abstract

A text analysis method is introduced which processes the unstructured information from document collection in order to support efficient text classification and information extraction. The Information extraction helps the system to discover useful knowledge for the users. The keyword based information search method fails due to the generation of many false positives. It leads to inaccurate generation of user expected result for many applications where context sense plays a major role while identifying the relevant features. The method developed can solve this problem by using combination of techniques from machine learning and natural language processing. The text contents are preprocessed and applied with entity tagging component. The proposed method performs accurate information extraction by improving precision and recall value. The automatic generation of rules provides an easy way to predict the effects of nutrient on malnutrition. The bottom-up scanning of key-value pairs helps to speed up the content finding to generate relevant items to the task. Our method can outperform other methods by eliminating sub tree which has an overestimating heuristic. Our approach will be useful to prevent errors generated through misinterpretation of available facts and improves the accuracy of information extraction in many text analysis applications.