The regular scientific seminar of department No. 1 was held at the Institute of Information Technology of ANAS.

Senior researcher of the institute Marziya Ismayilova introduced the presentation "Sentiment Analysis: Problems and Existing Approaches" and gave detailed information on the main issues, levels, classification methods, problems, etc.

The sentiment analysis is the extraction and recognition of the emotional content of the text from the text documents by using Natural Language Processing, statistics or machine learning techniques.

She noted that computer-based sentiment analysis was carried out with the presence of subjective texts on the Internet. The purpose of sentiment analysis is to find the ideas in the text and identify their features.

According to her, analysis of sentiment is a complex multidisciplinary issue that includes natural language processing, web analytics and machine learning.

She emphasized that classification of subjectivity is a method of classifying sentences, whether expressed or not, and that sentiment classification is a way to find out whether the text is in conflict or, in other words, positive or negative.

The speaker spoke about the level of sentiment analysis in terms of word, sentence, document and feature levels. The word-level sentiment analysis is based on a dictionary and a corpus. Support vector machine  (SVM), neural network,  Deep Learning,  Naive Bayes and other machine learning techniques were applied in the classification of sentiment analysis.

M.Ismayilova gave detailed information about the conference, symposiums and workshops on symmetry analysis, as well as researchers conducting research in this field.

 © All rights reserved. Citing to www.ict.az is necessary upon using news