Text analytics on Shiksha Reviews
- 1. “Text Analytics on Shiksha Reviews to Generate Meaningful Performance Indicators in Management Education” Shanmugapriya K Student, PGDM 2019-21, ISME Under the guidance of Dr. Shampa Nandi (HOD – PGDM) Professor Analytics and Marketing, ISME
- 2. Introduction The internet community in India has grown over 700 Million users in 2020 and is expected to grow in leaps and bounds. Performing text analytics with Educational data is very helpful in addressing issues, analyzing suggestions and finding solutions to the prevailing problems in the education sector. Feedback from the students on faculty, infrastructure, placement or management guide the institution to create a better learning experience. The present study is designed to investigate the learners’ feedback using Sentiment Analysis and Classification model and use Shiksha reviews on Management institutes of Bangalore.
- 3. Literature Review (1) “Sentiment Analysis for Education”: Sentiment Analysis is one of the widely used application of NLP, computational linguistics and text analytics which is also called as subjectivity analysis, opinion mining, and appraisal extraction. (Altrabsheh, Medhat et al., 2018) “University Ranking Prediction System by Analyzing Influential Global Performance Indicators”: This research paper built a data model which predicts and ranks universities by considering global university performance indicators as features. (Tabassum, Hasan, et al., 2017) “Role Of Sentiment Analysis In Education Sector In The Era Of Big Data: A Survey”: Even though big data analytics have been a hot topic in today’s world, the adoption of them in education sector has been slow. This paper discusses the different applications in which big data can be used. (Rao& Baglodi, et al., 2016 )
- 4. Literature Review (2) “Using Feedback Tags and Sentiment Analysis to Generate Sharable Learning Resources”: Sentiment Analysis can also be used to find the difference in perception of opinions and feedbacks from students. This paper investigates the results received from sentiment analysis combined with feedback tags. (Cummins, Burd, et al., 2010) “Do online reviews matter? — An empirical investigation of panel data”: This research considers reviews to be both influencing and influenced factors. Online reviews are endogenous in nature which changes the analysis. (Duan, Gu, et al., 2008) “Short Text Classification in Twitter to Improve Information Filtering”: As short texts do not provide sufficient word occurrences, traditional classification methods such as “Bag-Of-Words” have limitations. To address this problem, we propose to use a small set of domain-specific features extracted from the author’s profile and text. (Bharath Sriram, et.al, 2010)
- 5. Objectives of the study are- Performing Sentiment Analysis on reviews from shiksha.com and classifying the students’ reviews as positive and negative. Matching ratings and reviews and validating the honesty of the reviews. Predicting and building a ranking system of the Management institutes based on learners’ feedback.
- 7. Analysis Classifying Sentiments of the feedbacks: To classify the sentiments of the reviews into positive, negative and neutral, Vader Sentiment Intensity Analyzer is used.
- 8. Clustering Insights: • There are 1085 reviews which has an average of 4.1 overall rating. • 628 reviews with an average of 4.0 overall rating. • 718 reviews with an average of 3.5 overall rating.
- 9. Top Performers: MBA institutes (Colleges in the 2nd Cluster) • Regional College of Management • IIM Bangalore • IBS Bangalore • ISBR Bangalore • WeSchool Bangalore Third Best (Colleges present in the 0th Cluster) • Presidency University, Bangalore • DSCE Bangalore • XIME Bangalore • NMIMS SOM Bangalore Second Best Institues (Colleges present in the 1st Cluster) • East Point Group of Institutions • International Institute of Business Studies • SJB Institute of Technology • Mount Carmel College • ABBS Bangalore
- 10. Cluster 0 Cluster 1 Cluster 2
- 11. Limitations Sample size is quite small to build a ranking system.
- 12. Conclusion A sentiment intensity analyser model is built to classify the sentiments of the reviews. Clustering was done to group the colleges as Top Performers, Second Best Institutes and Third Best.