Identifying attitudes, behaviors and information pollution in epidemics using data mining and natural language processing techniques

Today, it is possible to see the elements of the information society in every aspect of life. This situation is clearly reflected in data production statistics. With 2012 figures, 2.5 quintillion bytes of data are generated daily in the world. All of the work such as processing and transferring big data at this scale is called Big Data. The fact that people who are not experts in their field present some things that are not based on any experience and that have not been passed through a filter of reason and logic, as if they were correct information, has led to the emergence of the concept called "information pollution". On the internet, which is an uncontrolled medium, it has become very difficult to access valuable and useful information as a result of people writing anything that has not been proven to be true and whose reliability is not questioned and whose source is uncertain and sharing it as if it were real information. Investigating the negative effects of this information pollution on health management and taking rational measures against these negative effects can facilitate health management. As the most recent example, unsourced information shared unconsciously or intentionally by non-specialists on social media channels during the Covid 19 pandemic has made epidemic management difficult. In line with the aforementioned reasons, trying algorithms to detect and reduce information pollution in this field and conducting analysis studies in this direction may be useful in terms of solving the problem. In our study, the effects of information shared on social media (Twitter) in terms of epidemic and health management were investigated using data mining techniques. The effects of information pollution on public health were tried to be analyzed with the data collected from internet-based social media through programming languages and their effects were presented.