MIS 2502: Text Mining and Sentiment Analysis
Text mining and sentiment analysis have become a very crucial part of data analytics due to the huge surge of blogs and forums as a result of Web 2.0. These sites can harbor anywhere from hundreds to thousands of people looking to be sold a new product or service. Companies all over the world are looking for gold mines like forums to get a feeling of what products customers like on the market, and what they would like to see in the future. These may be evaluations, judgments, or intended messages that are greatly valued by those looking to get consumer data and make sense of it. Text mining and sentiment analysis allows companies to get a feel for how consumers react based on the way things are written on forums and blogs.
Text sentiment mining seeks to find a positive, negative, or neutral feeling from a document or even on a more advanced level, feelings like happy or sad. Classification and cluster analysis go hand-in-hand with text and sentiment mining in that documents that have a specific sentiment are clustered together using computer software and algorithms to cluster different forums into groups based on feelings towards a certain topic. The algorithms created take text that was given a sentiment value(attribute) and classify them under a specific group(entity). Clustering helps to put groups of blogs and forums into something known as “hotspots” by using statistical formulas to create similar groups based on sentiments and make it easier to observe sentiment within a large set of data.
Depending on what type of program is used, those analyzing the data may see mathematical representations, visual representations, or both in the model used to obtain answers on certain hotspots. The University of California has a case study entitled, “Using text mining and sentiment analysis for online forums hotspot detection and forecast,” which specifically analyzes the use of cluster analysis and classification through text sentiment mining. Hotspots were determined each week of the year based on the number of visits to each site. Based on words and feelings alone, social media will be a driving factor in determining for companies which products will work, or should be modified or removed from the market on the whole.
References:
Li, Nan, and Desheng Dash Wu. “Using Text Mining and Sentiment Analysis for Online Forums Hotspot Detection and Forecast.” Decision Support Systems 48.2 (2010): 354-68. Decision Support Systems. 15 July 2008. Web. 1 May 2012. <http://www.cs.ucsb.edu/~nanli/publications/N.Li_2010_DSS.pdf>.
Pang, Bo, and Lillian Lee. “Opinion Mining and Sentiment Analysis.” Foundations and Trends® in Information Retrieval 2.1–2 (2008): 1-135. Open Mining and Sentiment Analysis. 2008. Web. 1 May 2012. <http://www.cse.iitb.ac.in/~pb/cs626-449-2009/prev-years-other-things-nlp/sentiment-analysis-opinion-mining-pang-lee-omsa-published.pdf>.