The goal is to summarise, compare, and recommend (with explanations) products based on features and associated sentiment mined from product reviews.
- Provided with a dataset of (6,000) reviews for two product domains (digital cameras and printers)
- Opinion mining – extract product features (feature sets are provided) and associated sentiment from the product reviews for each product
Functionality:
- Product summary – the table on previous slide represents the basic deliverable for this task, top marks going for novel/interesting product summaries.
- Product comparison – compare two or more products based on common/distinct features & associated sentiment.
- Product recommendation and explanations – for a given (target) product, provide recommendations of similar products to the target product. Provide explanations for each recommendations made – e.g. recommendation X has better sentiment for features A, B, and C compared to current product…