I have to do a project in which using the product description of any product at any e-commerce website and comparing it with all of the reviews of the product.I have to find out if the description of the product is false or not.
Abhishek Tanwar ,
I would like to clear some points in your question to see if I can be of assistance.
1) When you say negative of another, are you implying antonymy or is it based on a valuation on a scale(e.g. positive comment/negative comment), or other criteria?
2) What are the parameters of the project? I ask this since this is quite a complex project due to:
a) That the correlation between a statement and a fact can be derived from a review of the product(reviews do not obligate a customer to exclude subjective expressions that have nothing to do with the actual product performance)
b) Assumes that automated robots are not filling in the reviews
Let me know
Regards
If I understand your question. You want to evaluate the description (i.e the information provided by the retailer) about a product by considering the comments posted by users about that product.
You may extract aspects from the description and try to get the sentiment given by customers about each aspect if its positive than what's in the description could be true, and if it's not the description could be not true.
To do this you may use CNN as in https://github.com/soujanyaporia/aspect-extraction
Hi, if you are going to do polarity analysis (Positive, Negative and Neutral), you can use the Textblob library via Python to solve the problem. You can also get the emotional subjectivity rates of comments. I hope I understood the question correctly. Regards.
- Badges
- Science topic