Reputation analysis of brands in social media

Social networks are actively used nowadays and users can express their opinions by posting contents on these platforms. Social media posts about a brand affects the opinions of other users and therefore the reputation of the related brand. Thus, most of the brands follow posts about themselves on social media platforms and accordingly take actions. Within the scope of this thesis, it is aimed to develop an information system that enables brands to analyse their reputation in social media and manage their relationship with consumers in social networks. For the aimed system, software with multi-layer architecture has been developed in order to collect the posts about the brands. In the process of classification of the collected content as positive or negative, machine learning methods were applied, since using machine learning for the classification of contents is faster and inexpensive compared to human power. The Naive Bayes classification model with 90% F-criterion score was chosen among the machine learning models designed. Lastly, Brand Reputation Management System (BRMS), web application in MVC pattern, was developed to present the automatically classified posts and their analyses. The user, registered to the system for a brand, can view the posts and analysis about the related brand by using BRMS. Additionally, the registered user can take some actions such as responding, re-sharing, and like to a post via BRMS. In future work, it is aimed to integrate the current system (BRMS) to many social media platforms and make the system more comprehensive. Keywords: Machine Learning, Text Mining, Brand Reputation, Social Media, Information Systems