Využití analýzy sentimentu trhu v B2C organizacích

Sentiment analysis is a concept which is gaining in importance in connection with the growing data volume and data diversity. Increasingly, organizations are forced to obtain new information from unstructured data coming from new media and other channels.This diploma thesis deals with the topic of sentiment analysis primarily from the viewpoint of its practical use in organizations. The aim is to define areas, where the functioning of an organization can be advanced by sentiment analysis and exactly compare the available commercial tools that allow this analysis in Czech language.The first part of the thesis defines key concepts, the relation between the topic and information management; and presents the concepts of big data, machine learning and practicable ways of the sentiment analysis application. Subsequently, the existing research are charted by sectors, where this analysis can be employed.The next part of the thesis brings in specific commercial software solutions that make sentiment analysis available in Czech language. These solutions frequently operate as a kind of a black box, and it is difficult to compare them reciprocally and accurately with respect to their ability for correct data classification. Therefore, their functions are displayed, and their ability for correct data classification is measured on the created dataset of user reviews. Besides other things, the characteristics of the macro-averaged F1 measure or the root-mean-square error (RMSE) are monitored.At the end of the thesis, potential distortions are discussed, and the principal recommendations, which should be considered before the implementation of sentiment analysis tools in organizations, are summarized.