Business organizations have been adopting different strategies to impress upon their customers and attract them towards their products and services. On the other hand, the opinions of the customers gathered through customer feedbacks have been a great source of information for companies to evolve business intelligence to rightly place their products and services to meet the ever-changing customer requirements. In this work, we present a new approach to integrate customers' opinions into the traditional data warehouse model. We have taken Twitter as the data source for this experiment. First, we have built a system which can be used for opinion analysis on a product or a service. The second process is to model the opinion table so obtained as a dimensional table and to integrate it with a central data warehouse schema so that reports can be generated on demand. Furthermore, we have shown how business intelligence can be elicited from online product reviews by using computational intelligence technique like rough set base data analysis.
TopOpinion Mining
Opinion mining popularly known as sentiment analysis has been evolved as an interesting research area in recent years. Most research has focused on analyzing the content of either product review (Dave, Lawrence & Pennock (2003) or movie reviews(Pang et al, 2004). Furthermore sentiment analysis has been extended to other domains such as news, blogs and debates also. Sentiment analysis or opinion mining refers to the application of natural language processing, computational linguistics and text analytics to identify and extract subjective information implicit in source materials. It is an excellent means for handling many business intelligence tasks as it describes sentiment analysis as a process that categorizes a body of textual information to determine feelings, attitudes and emotions towards a particular issue or object. Generally speaking, sentiment analysis aims to determine the attitude of a speaker or a writer with respect to some topic or the overall contextual polarity of a document. The attitude may be one’s judgment, assessment, sensitiveness or emotional state of the author when commenting about a subject. Opinion mining deals with detection and classification of sentiments in a text (Bifet & Frank, 2010). Sentiment analysis research achieves two things: i) identifying the polarity of the underlying subject in the text, and ii) determining the strength of the polarity (severity or intensity). Generally, the sentiment polarity is classified as positive, negative or neutral classes and the strength of polarity is expressed in numeric figures. Keywords like dazzling, brilliant, phenomenal, excellent, fantastic, spectacular, cool, awesome, thrilling obviously expresses a favorable context of the subject while keywords like terrible, awful, worst, horrible, stupid, waste express unfavorable sentiment. Sentiment elicitation is done at different levels focusing on either single words, phrases, complete sentences or a complete document by adopting techniques such as unigrams, bi- grams, N-Grams, and opinion words.