ISSN 2319-2720 Volume 2, No.1, January – March 2013 International Journal of Computing, Communications and Networking Available Online at http://warse.org/pdfs/2013/ijccn08212013.pdf
Arvind Kumar Jain et al, International Journal of Computing, Communications and Networking, 2(1), January – March 2013, 36 - 40
Analysis and Implementation of Sentiment Classification Using Lexical POS Markers 1
1
2
Arvind Kumar Jain , Yogadhar Pandey M.Tech, SIRT Bhopal (MP),India, arvindjain98@gmail.com 2 SIRT Bhopal (MP), India, p_yogadhar@yahoo.co.in during the formulation of linguistic theories and frame-works [1, 2]. Languages can be represented by their two principal components, a lexicon and a grammar. A language's lexicon contains the legal combination of letters or symbols which represent a meaning. Lexicons are not static; new words are added with new meanings, old words may change or take on a new meaning, and words may become obsolete and eventually work their way out of the lexicon altogether. The basic unit of a lexicon is a "lexeme", a sequence of one or more words. Lexemes have syntactic properties, which define their structural relation to the grammar of the language, and semantic properties which define the 'meaning' of the lexeme symbolically[2]. It has been shown that syntax and semantics are closely related in technical and scientific sublanguages. The way in which a word combines syntactically and semantically with other words can be used as a basis for a classification system. Sentiment Lexicons The semantic orientation of a term indicates its capacity for carrying positive or negative evaluative value. It is possible for this information to exist a priori with relative independence from the context it may appear, as seen on words such as “excellent” or “terrible”. For this reason knowledge of such terms can be a useful when identifying sentiment and is a motivation for the development of collections of opinionated terms into a sentiment lexicon. Sentiment lexicons exist as manually annotated databases such as the General Enquirer mapping terms in the English language into semantic categories, including sentiment orientation. Initially compiled to assist research on social studies, it has proven useful on opinion mining research and is regarded as a highly accurate lexicon used as a baseline for comparisons [3]. Other ad-hoc manual resources were generated for specific research. However the collection and annotation of a large sized lexicon is a expensive and time consuming task and has motivated research in automated methods that leverage existing language resources to build or expand existing lexicons. Sentiment analysis The task of identifying positive and negative opinions, emotions and evaluations. Most work on
ABSTRACT Natural language processing has attracted many researchers as the amount of information available on the internet and other media is increasing exponentially. So we need computer interference to process that enormous data.NLP does exactly the same by converting the natural language into something to be understood by computer to process. Though it has a lot of branches I am working in sentiment analysis of text. By this we mean a method to find what the reader is thinking when he/she was writing a particular text. This helps in tracking possible suicidal case, terrorist attacks and also in finding the orientation of a particular product and changing ourselves according to customer reviews. Polarity analysis is a subset of sentiment analysis when we find the polarity or orientation of a particular word and thus finding the overall polarity of a sentence, this is useful in finding whether a particular review of a product/movie/hotel or anything else is good or not [1]. With the ever increasing number of websites available free to post reviews today’s customer are smart to follow the reviews before purchasing a product. While many review sites, such as Epinions, CNet and Amazon, help reviewers quantify the positivity of their comments, sentiment classification can still play an important role in classifying documents that do not have explicit ratings. Key words: analysis, negative, polarity, positive, Sentiment INTRODUCTION The recent success of data-driven approaches in NLP has raised important questions as to what role linguistics must now seek to play in further advancing the field. Perhaps, it is also time to pose the same question from the other direction: As to how NLP techniques can help linguists make informed decisions? And how can the advances made in one field be applied to the other? Although, there has been some work on incorporating NLP techniques for linguistic field-work and language documentation (Bird, 2009), the wider use of NLP in linguistic studies is still fairly limited. However, it is possible to deepen the engagement between the two fields in a number of possible areas, and gain new insights even 36 @ 2012, IJCCN All Rights Reserved