I'm working on some aspcet based sentiment analysis project applied on some tweets or reviews, so far I've applied all the known preprocessing methods on my training dataset which is an XML file that you can find on THIS LINK.
What I wanna do next is extracting related words to some aspects, so I can pass them through my classifier to decide whether it's positive or negative, here's an example to make it clearer:
let's take the first example of my dataset (first tweet);
< sentence id="2339"> < text>I charge it at night and skip taking the cord with me because of the good battery life.< /text> < aspectTerms> < aspectTerm term="cord" polarity="neutral" from="41" to="45"/> < aspectTerm term="battery life" polarity="positive" from="74" to="86"/> < /aspectTerms> < /sentence>
so as you can see in this example there are two terms that I should classify the sentiment expressed about, my problem isn't the extraction of the terms which goes back to the ''aspect terms extraction'' famous problimatic, but it's how to decide what are the words associated to those terms, in the example I presented; ['charge', 'night', 'good'] are associated to the term 'battery life', while ['skip', 'taking'] are associated to the term 'cord'. In other words, I would love to have extract from one sentence (tweet) the words related to my Aspect terms separated from each other.
can anyone give me a hint or some oriontation about how to figure this out?