Generation of Regular Expressions for Large Clinical Dataset using NLP and Machine Learning Techniques

Abstract

Machine learning supervised classification plays major role in large text classification. In health care data it contributes since couple of years for generating the privacy and security. Such kind of electric records might take large data on storage devices, so it needs to optimize with some processing techniques. The Natural Language Processing (NLP) features to represent text as a vector appropriate for use in machine learning algorithms. To describe text as just a vector suitable for use in machine learning algorithms, text classification techniques frequently rely on dictionary counters including bag-of-words as well NLP techniques. As an alternative feature set, editable search queries are proposed in this work. The suggested model utilizes the Smith Waterman (SW) optimization created by Pair -wise alignment to create a set of pattern matching characteristics based on labeled textual information or to train a later part classifier. The Naive Bayes (NB) supervised learning algorithm has used for classification. While a correlation of the special groups and traditional text characteristics shows that a classification algorithm does not make better decisions utilizing generated features, the produced factors can catch patterns that cannot be identified with collective expertise. As a result, it is significantly boost a classifier that combines conventional methods for generated characteristics. In the experimental analysis we evaluate the proposed system results with various existing systems that our system demonstrates effective results than traditional approaches.