Yelp-Dataset-Challenge
Task 1
Recommending business to users
Motivation
Our aim is to recommend business restaurants to users using two approaches and evaluate both the methods in order to determine which is a better approach to recommend business with data that we have generated.
Two approaches used for recommendation
-
Collaborative filtering
- Memory Based approach
- User based
- Item based
- Model Based approach
- SVD Decomposition
- Memory Based approach
-
Hybrid approach : Content based + Collaborative approach
Task 2
Keyphrase Extraction
Motivation
Our aim is to extract relevant key phrases from reviews of businesses and carry out indexation and retrieval of Business reviews based on selected tags.
Users may want to read reviews related to specific features/tags, say, some users may want to read reviews that contain information about particular dishes in restaurants, or the service, or prices. Our aim is to extract relevant and important features associated with businesses from the business reviews given by users. Using these tags, users can identify important features of a business without having any prior knowledge of the business and also filter reviews to make an informed decision.
We also designed and developed a prototype for reviews retrieval on Yelp Dataset. Make sure to go through to the demo video to understand how our system funtions. You can find the video here.
Group 3 Team Members:
- Arpit Shah: arpishah@iu.edu
- Neha Pai: nrpai@iu.edu
- Nikita Bafna: nibafna@iu.edu