Academic journal article Journal of Digital Information Management

Detecting Human Sentiment from Text Using a Proximity-Based Approach

Academic journal article Journal of Digital Information Management

Detecting Human Sentiment from Text Using a Proximity-Based Approach

Article excerpt

I. Introduction

The amount of textual data accumulated each day by various businesses, scientific, and governmental organizations around the world is daunting. Success in the development of statistical natural language processing (NLP) has led to improvements in fundamental text analysis such as part-of-speech (POS) tagging, phrase chunking, dependency analysis and parsing. Using these components as fundamental building blocks, many NLP researchers have become interested in analyzing text "semantically" or "contextually". For example, named entity tagging, semantic role tagging and discourse parsing are being investigated in the NLP fields. This move towards taking contextual or semantic information into account has occurred in application areas of NLP such as text classification, text summarization, information retrieval and question answering. Even with text classification tasks, one of the traditional NLP tasks, the target classes have recently been diversified from topics such as 'sports' and 'economics' to the contents of texts such as 'polarity' or 'subjectivity'. This thus calls for methods for sentiment classification or sentiment analysis [1, 2].

Sentiment analysis is an emerging field, concerned with the detection of human emotions from textual data. Sentiment analysis can be used for grouping search engine results, analyzing news content, reviews for books, movies, sports, etc and different types of blogs, social networks and web forums. It is quite difficult for the computer to pull out the tone and meaning of a document automatically because people express things in many different ways. In this paper, we propose a novel approach to sentiment analysis. We call it proximity-based sentiment analysis. We applied our proximity-based sentiment analysis technique to the problem of analysis of movie reviews. Accurate automatic prediction of how much an audience is going to enjoy a movie is still a challenging research question. In this paper, we systematically study the accuracy of different classifiers in the movie review domain, where the input to the classifiers are proximity based features. We also show the performance on other domains, namely drug reviews and music reviews.

The paper is organized as follows: In section II we present a background to the problem we address. Section III presents our methodology. Section IV presents our results and discussion. The paper is concluded in Section V.

2. Background

Early research on sentiment analysis [3,4] used models inspired by cognitive linguistics or manual or semi-manual construction of discriminant-word lexicons [5]. In [6] text classification was performed based on the sentiments or genre. As discussed in [7], researchers doing sentiment analysis seem to focus on specific tasks, such as finding the sentiments of words [8], subjective expressions [9], subjective sentences [10] and topics [11], or extracting the sources of opinions [12]. Pang and Lee [10] proposed a machine-learning method to improve the effectiveness of sentiment classification algorithms. DasGupta and Ng [13] proposed a text clustering algorithm based on the author's mood.

A number of researchers investigated hate groups using sentiment analysis techniques. For instance, hate crime on the internet was studied in [14]; the use of the Internet by terrorists, extremists and activists was investigated in [15, 16]. A study on US hate group web sites can be found in [17, 18]. In [19], Burris et al discussed the importance of analyzing web forums and chat-room content and use social network analysis to examine the inter-organizational structure of white supremacist movements. In [20], customer feedback was used to assign sentiments to documents using a 4-point scale. Similar work was proposed in [21] for movie review analysis. Earlier, Pang, Lee and Vaithyanathan [22] performed sentiment analysis on movie review data. While Dave, Lawrence, Pennock [23] used a number of machine learning techniques to analyze product review data. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.