Organizational Online Reputation Measurement Through Natural Language Processing and Sentiment Analysis Techniques

The set of perceptions held by various groups based on history and expectations constitutes the reputation of organizations. There are multiple correct measurements of reputation since no general definition of the concept has been reached. ORM (Online Reputation Monitoring-management) systems oversee this measurement and have a sentiment analysis component to perform this task. The literature presents different frameworks or methodologies for measurement developed by academia and industry. These proposals’ common objective is to measure online reputation based on the opinions expressed by individuals close to the organization. In the absence of an automatic ORM system, it is necessary to perform this task manually within a company by a person; this can generate operational errors, delay processes, and make scalability impossible to increase the number of items reviewed (news, comments). These drawbacks can be mitigated by automating the measurement of a client’s online reputation. This paper contains the development of three methodologies from the literature to explore online reputation measurement starting from Twitter and Google News information sources. The implementation results conclude that the POS-Tagger elimination methodology generates the best result compared to the coded methodologies.