Can ontology be used or dictionary is to be used for analysis? I am doing my research with python and nltk components. I have different types of datasets like movies, electronic datasets, baby products etc. How should I use ontology here? Can anybody suggest and explain a detailed explanation? How can I identify noun phrases? Next time I will have to use feature extraction.
Using ontologies:
Ontology-based sentiment analysis of twitter posts http://dx.doi.org/10.1016/j.eswa.2013.01.001
or
AN ONTOLOGY BASED SENTIMENT ANALYSIS - A Case Study
http://www.diva-portal.org/smash/get/diva2:552748/FULLTEXT01.pdf
or
From Tags to Emotions: Ontology-driven Sentiment Analysis in the Social Semantic Web
http://ceur-ws.org/Vol-771/paper2.pdf
Using dictionaries:
Corpus-Based Dictionaries for Sentiment Analysis of Specialized Vocabularies
http://www.kenbenoit.net/pdfs/NDATAD2013/Rice-Zorn-LSE-V01.pdf
http://fjavieralba.com/basic-sentiment-analysis-with-python.html
Hopefully, this helps.
@Nithya
Ontologies can serve as the representation of the domain of interest. So, if you want to perform sentiment analysis on hotel reviews e.g., you need to have an appropriate ontology describing the various notions (entities-classes) and their properties (features), so that your application can then "understand" that a certain piece of text refers to class A and certain properties of this class.
@Michael
Thanks a lot for referring to my work "Ontology-based sentiment analysis of twitter posts" ! :P
By the way, an excellent introductory text to sentiment analysis is the book "Sentiment Analysis and Opinion Mining" (available online at http://www.cs.uic.edu/~liub/FBS/SentimentAnalysis-and-OpinionMining.pdf) by Prof. Bing Liu (http://www.cs.uic.edu/~liub/). Personally, it helped me a lot to understand the basic notions behind sentiment analysis.
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators