Swarm Optimized Opinion Classification Model for Policy Assessment
- 1. Department of Computer Science & Engineering, Delhi Technological University, Delhi, India.
- 1. Publisher
A government policy is a scheme launched by the governing body of a nation for the welfare of a particular section of the society or the entire public in general. The impact of such a policy can hence only be determined by the response from its target group. The evaluation of these schemes is often challenging, due to the inability of the government body or organization to collect unfiltered and unbiased feedback from the entire population. The aforementioned task may require a large amount of effort, considerable time and in-depth knowledge of advanced technology. However, with the advent of the information era, it is possible to analyze the sentiments of the public using negligible resources. The internet is rich in freely available unused and unstructured data that can be exploited efficiently for various purposes. One such application is opinion mining which allows the user to extract data from social media websites and categorize it into pre-defined classes. This paper is an attempt to assess one of the most important and current government initiatives- “Digital India”, through public sentiments. Digital India is a program launched by the Prime Minister of India to transform the country into a technologically advanced and digitally connected nation. This research work corroborates the use of swarm intelligence or nature-inspired algorithms for feature subset selection during opinion mining, as it results in a substantial reduction in the number of features (and consequently a lesser computation time for model training) and increase in the classification accuracy of the model. Therefore, the aim of this study is to analyze public opinion on “Digital India” campaign to ascertain the success (or failure) of the mission, while at the same time, determine the most suited model for automated evaluation of any government policy in the future.
- Is cited by
- Journal article: 2249-8958 (ISSN)
- Retrieval Number