Exploring the use of Machine Learning for Highly Accurate Text-based Information Retrieval System


  • Chandrashekhar Himmatrao Sawarkar
  • Pramod N. Mulkalwar


Information retrieval has been a pillar-stone for today’s digital age, wherein every online entity expects correct and fast information. More than 80% of all searches made on search engines are text-based and thus having an accurate text-based information retrieval system is a must for today’s corporations. Text-based retrieval systems range from simple query processing, to complex elastic search-based systems. The decision of algorithm selection for retrieval systems depends on the application for which the system is designed. While systems like chat-bots require highly-complex machine learning-based retrieval systems, some systems like intranet-based searches give high accuracy with simple query-processing. In this work, we propose the design of a machine learning-based hybrid information retrieval system which adapts itself as per the application, and provides solutions that result in highly accurate information retrieval. We tested the proposed algorithm under different datasets, and found it to be accurate with lesser response time as compared to some of the state-of-the-art systems.