Semantic Outlier Detection

Between the dawn of the Internet through year 2003, there were just a few dozens exabytes of information on the Web. Today, that much information is created weekly. The opportunity to capture the opinions of the general public about social events, political movements, company strategies, marketing campaigns, and product preferences has raised increasing interest both in the scientific community, for the exciting open challenges, and in the business world, for the remarkable fallouts in social media marketing and financial forecast. Keeping up with the ever-growing amount of unstructured information on the Web, however, is a formidable task. Unlike standard statistical approaches, sentic computing relies on a vector space model of affective common-sense knowledge to work with natural language at conceptlevel. The well-known noisiness of common-sense data sources, however, is a major factor in jeopardizing the efficiency of analogical reasoning in the vector space. In this work, it is explored how least absolute deviations can aid semantic outlier detection and, hence, enhance concept-level opinion mining.