Thesis work Title: A Novel Framework For Recommendations Based On Analyzing Sentiments From Textual Reviews
Author: Abhishek kumar
Course: M.Tech (CSE)
College: CGC-COE, Landran, Mohali (PB)
Month/Year of Submission: Sept. 2017
From the beginning of 1990s, Recommender systems have been an integral part of Research to overcome information overload issue. Recommender Systems are everywhere, from E-Commerce websites to Academic Research. Every person is having unique taste and preferences for the objects(living/non-living) they interact with. Based on that taste and preferences what might be the similar objects user might find interesting and might be recommended to him/her. If person enjoys reading web development books, then based on his/her preferences what might be other good web development books, he/she might like if recommended. There are various approaches for the design of Recommender system; and Sentimental Analysis is among one of them. For most of Sentiment Analysis based Recommender Systems’ the core of whole model relates closely with the Rating Prediction Task. The whole Recommender System depends on computed/predicted ratings directly or indirectly. In this thesis work, NRPS Model was proposed. NRPS is a Rating prediction system, which predicts the Ratings by mining the sentimental information from social users’ reviews. For experiments, a real-world dataset was chosen from Yelp(ii). From the users’ generated reviews, product features were extracted via LDA algorithm and Sentimental dictionaries were constructed for the computation of ratings scores. ANN was used for reviews’ ratings scores prediction. The proposed model was implemented via 3 technologies MatLab(iii), Java(iv) and PHP(v). Proposed Rating Prediction model has been compared with RPS and SVMRPS Models. For performance analysis, two performance evaluation metrics—RMSE and MAE were used. Experimental Results shows the significant improvement over the RPS  and SVMRPS Models, over a real world dataset.
I would like to thank Dr. Manish Mahajan (HOD, CGC-COE), both of my guides Mr. Ranjeet Singh and Ms. Deepika Sood; CGC College of Engineering, Landran, Coordinator, for their kind support.I also owe my sincerest gratitude towards Ms. Dapinder Kaur (Class Coordinator), Department of CSE CGC-COE Landran, Mohali for her pearls of wisdom, during the course of this research. I also owe a gratitude towards my classmates especially Neeraj Baatish, Harpreet Kaur, faculty members (CGC-COE, CSE Department)—Mr. Jagbir Singh Gill, Mr. Gaurav Goel and Mr. Tejpal Prasher; for their guidance and support during my M.Tech course.
 X. Lei, X. Qian, and G. Zhao, “Rating Prediction Based on Social Sentiment from Textual Reviews,” IEEE Trans. Multimed., vol. 18, no. 9, pp. 1910–1921, 2016.
 A. Kumar, R. Singh, and R. Saini, “SVMRPS — Support Vector Machine based Rating Prediction System,” 2017.