07-25-2016 08:38 AM
I am new to using rapid miner for sentiment analysis/text mining. I watched couple of videos. I used the tokenize, filter, stop words etc.
I have customer id, review date, year of review (Computed from 2nd), customer location (From web scaped data) and the customer review. The field name in Excel is Customer_reviews.When I import the excel, the customer_review comes up as polynomial.
I also get all the positive/negative word list as one word in rapidminer, where as when I look it up in Notepad++, it is all in a separted lines.
How do I use the customer_review as a text to arrive at sentiments using postive/negative word list?
Could some one point me the right direction?
Thanks for your time.
Solved! Go to Solution.
07-25-2016 09:41 AM
Check out the two articles here around sentiment analysis
They will provide you some ideas on how to sentiment analysis in RapidMiner.
Also there are two extensions on Rapidminer that you can leverage
They are available via marketplace, Rosette and Aylien.
08-04-2016 06:19 AM
Through our Text Analysis by AYLIEN Extension we provide both Document-level Sentiment Analysis and Aspect-based Sentiment Analysis which works really nicely on customer reviews.
Document-level Sentiment gives you the overall polarity of a customer review while Aspect-based sentiment identifies certain aspects mentioned in a review and determines the sentiment towards those aspects.
You can read more about it here. http://blog.aylien.com/post/148356512853/aspect-based-sentiment-analysis-is-now-available
03-10-2017 12:05 PM
I'm trying to use AYLIEN extension to conduct an Aspect-based Sentiment Analysis.
There is a parameter called "domain" in which I have to choose between the following options: "cars", "hotels", "airlines" or "restaurants". My question is whether this extension is limited to only those domains. If not, which of them can we use as default?
Thanks in advance.
03-10-2017 12:31 PM
the aspect based ones are w.r.t these aspects. Simply choose the generic Extract Sentiment to get a general one.