A Practical Guide to Sentiment Analysis

ISBN-13: 9783319856483 / Angielski / Miękka / 2018 / 196 str.

Erik Cambria; Dipankar Das; Sivaji Bandyopadhyay

Spis treści

Preface

Affective Computing and Sentiment Analysis

Many Facets of Sentiment Analysis

Reflections on Sentiment/Opinion Analysis

Challenges in Sentiment Analysis

Sentiment Resources: Lexicons and Datasets

Generative Models for Sentiment Analysis and Opinion Mining

Social Media Summarization.- Deception Detection and Opinion Spam

Concept-Level Sentiment Analysis with SenticNet

Index.

Blurb

This edited work presents studies and discussions that clarify the challenges and opportunities of sentiment analysis research. While sentiment analysis research has become very popular in the past ten years, most companies and researchers still approach it simply as a polarity detection problem. In reality, sentiment analysis is a ‘suitcase problem’ that requires tackling many natural language processing subtasks, including microtext analysis, sarcasm detection, anaphora resolution, subjectivity detection and aspect extraction. 

 In this book, the authors propose an overview of the main issues and challenges associated with current sentiment analysis research and provide some insights on practical tools and techniques that can be exploited to both advance the state of the art in all sentiment analysis subtasks and explore new areas in the same context. Readers will discover sentiment mining techniques that can be exploited for the creation and automated upkeep of review and opinion aggregation websites, in which opinionated text and videos are continuously gathered from the Web and not restricted to just product reviews, but also to wider topics such as political issues and brand perception.

The book also enables researchers to see how affective computing and sentiment analysis have a great potential as a sub-component technology for other systems. They can enhance the capabilities of customer relationship management and recommendation systems allowing, for example, to find out which features customers are particularly happy about or to exclude from the recommendations items that have received very negative feedbacks. Similarly, they can be exploited for affective tutoring and affective entertainment or for troll filtering and spam detection in online social communication.