| Biography | |
|---|---|
 Prof. Alexander Gelbukh Mexican Academy of Sciences, Mexico |
|
| Title: Novel Techniques for Sentiment Analysis and Opinion Mining | |
| Abstract:
Opinion mining is a rapidly growing research area that aims on leveraging the huge amount of user-contributed contents in Internet for collecting data, opinions, and suggestions on the quality of as many of products and services as possible, ideally on almost every product or service on the market. In the context of business intelligence and decision-making support systems, this information is used by companies to improve their products, to develop new products, to distribute their products by market niches, and to improve targeted advertising campaigns according to the preferences of different user groups. In the context of recommender systems, this information is used by the consumers to find better and more suitable products, with which other users with similar needs and preferences have had positive experience, and thus to make more informed purchasing decisions. Indirectly, through a “natural selection” mechanism, this also leads to improvement in the average quality of products and services available in the market: bad products have less chance to be sold, and companies that cannot improve their products exit the market. Technically, opinion mining includes, as the most important steps, sentiment analysis (to find whether a review or social network post expresses positive or negative opinion about the product), author profiling (to determine the characteristics of the author of the opinion, such as psychological type), and deception detection (to filter out fake opinions). In this keynote talk, I will describe the technique our group has developed as contribution to these important tasks. In particular, I will present deep-learning architectures we have developed for these tasks.
| |
| Biography: Member of the Mexican Academy of Sciences National Researcher of Mexico (memeber of SNI) Excelence Level 2 Ex-president of the Mexican Society of Artificial Intelligence h-index: Google Scholar: 34, Scopus: 15, ISI WoS: 10. Erdős number: 4 Author of more than 500 publications | |
