Combining Knowledge and Data Mining to Understand Sentiment - A Practical Assessment of Approaches

About this paper

In our fast-paced world, consumer opinion can make or break your company's products or services virtually overnight. To understand consumer sentiment and respond quickly to opinions, you must continually monitor and evaluate relevant Web content. Most companies rely on automated techniques to help them quickly characterize the sentiment of documents in a variety of domains. This white paper compares two different approaches to sentiment analysis and illustrates why combining domain-specific linguistic rules with data mining methods improves both the effectiveness of models and the efficiency of model builders.