Geek Answers Handbook

Email Natural Language Processing Algorithm

One simple question (but I did not find an explicit answer in the NLP materials that I read that I am very new):

I want to classify emails with probability according to certain mood parameters. Is there an NLP package specifically dedicated to this? Is there an obvious starting point in the literature I'm starting to read?

For example, if I have a short letter like "Hi, I'm not very impressed with your last email address, you said that the order amount will be only $ 15.95! Regards, Tom," then he can get 8/10 for Disappointment and 0/10 for Happiness.

The actual mood list is not that important, but a short list of generally positive and generally negative moods would be helpful.

Thanks in advance!

- Trindaz on Fedang #NLP

+6
nlp
source share
3 answers

You can do this with a number of different NLP tools, but I don’t know that it is ready out of the box. Perhaps the easiest way to start is with LingPipe (java), and you can use their very good mood analysis guide . You can also use NLTK if python is more suitable for you. There are some good blog posts on Streamhacker that describe how you would use Naive Bayes to implement it.

+7
source share

Check the AlchemyAPI for mood analysis tools and scikit-learn or any other open computer learning library for the classifier.

0
source share

if you do not decide to code the implementation, you can also classify data with another tool. A google api view may be an alternative.

In any case, you will need some tagged data and do the preprocessing. But if you use a tool that can help you get more accurate accuracy.

0
source share

More articles:


All Articles