How to spot product features in an English sentence - nlp
I am trying to detect the features (eg: screen, processing speed) of a product (eg: mobile, respectively) in an English sentence. To do this, my approach is that in a paragraph (which speaks about a product) containing many sentences, words (other than words like pronouns or words of feeling like good, bad, etc., which I keep in a file ) that most often appear the features of this product and therefore I rank by their frequency and their distance using the words of feeling and take the top of them.
However, this is not very efficient. Can anyone suggest any other and better approach for detecting words that are features of a product?
source to share
A huge amount of research has been conducted in this area. Start by reading Bing Liu's original work (Liu 2004, Liu 2005) in this area.
One popular technique is using the Dependency Graph using Stanford CodeNLP. You can make rules like App-related Noun ( NN ) ( JJ ) using the nsubj dependency . 5-10 of these rules would be enough for a basic system.
The prior art in this area uses a sequence labeling (CRF / HMM) approach to tag each word, regardless of whether it is a functional term or not. However, you need a good amount of tagged data to do this. Check out the latest work in Aspect Based Mood Analysis.
Resources for your help:
http://alt.qcri.org/semeval2015/task12/
http://www.aueb.gr/users/ion/docs/pavlopoulos_phd_thesis.pdf
http://www.aclweb.org/anthology/S14-2004
source to share