Sentiment Analysis as a Restricted NLP Problem

Introduction

With the emergence of WWW and the Internet, the interest of social media has increased tremendously over the past few years. This new wave of social media has generated a boundless amount of data which contains the emotions, feelings, sentiments or opinions of the users. This abundant data on the web is in the form of micro-blogs, web journals, posts, comments, audits and reviews in the Natural Language. The scientific communities and business world are utilizing this user opinionated data accessible on various social media sites to gather, process and extract the learning through natural language processing. In this way, there is a need to detect and distinguish the sentiments, attitudes, emotions and opinions of the users from the user’s generated content. Sentiment Analysis is the process which aids to recognize and classify the emotions and opinions of users in the communicated information, in order to determine whether the opinion of the user towards a specific service or product is positive, negative or neutral through NLP, computational linguistics and text analysis. While this user opinionated data is intended to be useful, the bulk of this data requires preprocessing and text mining techniques for the evaluation of sentiments from the text written in natural language. Sentiment Analysis permits organizations to trace their brand reception and popularity, enquire about new product perception and anticipation by the consumers, improve customer relation models, enquire company reputation in the eyes of customers and to track the stock market. According to the Local consumer review survey (Bloem, 2017), 84 percent of the total people trust online reviews as much as a personal recommendation given to them. Thus, it is important to mine online reviews to determine the hidden sentiments behind them.

According to Techopedia (2014), Sentiment Analysis is defined as “a type of data mining that measures the inclination of people's opinions through NLP, computational linguistics and text analysis, which are used to extract and analyze subjective information from the Web- mostly social media and similar sources”. The analyzed data measures the consumer's experiences and opinions towards the products, services or proposed schemes and discloses the contextual orientation of the content. Sentiment analysis encounters many challenges due to its analysis process. These challenges become hindrances in examining the precise significance of sentiments and identifying the sentiment polarity. Some of the common challenges faced by sentiment analysis include difficulties in feature extraction, increased complexity in analyzing label opinionated data, the complication in analysis of other regional languages, requirements of world knowledge, increased domain dependency etc. Unfortunately, sentiment analysis also experiences various difficulties due to the sophisticated nature of the natural language that is being used in the user opinionated data. Some of these issues are generated by NLP overheads like colloquial words, coreference resolution, word sense disambiguation and so on. These issues add more difficulty to the process of sentiment analysis and emphasize that sentiment analysis is a restricted NLP problem. Different algorithms have been applied to analyze the sentiments of the user-generated data. The techniques applied to the user-generated data ranges from statistical to knowledge-based techniques. Even hybrid techniques have been used for the sentiment analysis. Various algorithms, as discussed above, have been employed by sentiment analysis to provide good results, but they have their own limitations in providing high accuracy. It is found from the literature that deep learning methodologies are being used for extracting knowledge from huge amounts of content to reveal useful information and hidden sentiments. Many researchers have explored sentiment analysis from various perspectives but none of the work has focused on explaining sentiment analysis as a restricted NLP problem.

Thus, this chapter presents an overview of Sentiment analysis, which is followed by the related work in section 2, then the detailed description of generally employed methodologies and techniques in Sentiment analysis are discussed in Section 3. Section 4 explains the applications of Sentiment Analysis. Section 5 describes the challenges faced by the Sentiment Analysis and then the challenges relevant to NLP are discussed in Section 6. Section 7 explores the solutions and recommendations to resolve the challenges and in the next section, some future research directions have been explored.