Unsupervised Sentiment Analysis on product reviews

Question

I would like to perform an unsupervised sentiment analysis on the reviews posted by customers on different product web-page. Most of the online resources use supervised methods and the examples/tutorials always have a labelled training data-set.

My objective is not to just deduce the polarity of the review but also do content/subjective analysis. Any online implementation or suggest high level approach for this ?

I'm voting to close this question as off-topic because this question belongs to datascience.SE — Matthieu Brucher, Dec 17 '18 at 11:21
This question is off topic for stack overflow, but it is on topic for datasxience.stackexchange.com — Matthieu Brucher, Dec 18 '18 at 08:02

score 0 · Answer 1 · answered Dec 18 '18 at 16:38

0

One unsupervised approach you can take is word vector embedding along with k-means clustering. Here's an example in Python http://brandonrose.org/clustering

answered Dec 18 '18 at 16:38

Adnan S

1,852
1
14
19

score 0 · Answer 2 · answered Dec 26 '21 at 15:37

I know this is 3 years later but since there is no accepted answer imma answer it.

My objective is not to just deduce the polarity of the review but also do content/subjective analysis. You can do just this with the help of a library called TextBlob. Here is a link to an article where it is mentioned how to get polarity and subjectivity from a string using TextBlob.

Unsupervised Sentiment Analysis on product reviews

2 Answers2