Developing a model of association rules with machine learning in determining user habits on social media

The object of the research is the habits of social media users. The problem in this research is that there is a lot of data spread across social media platforms ranging from image data, text data to audio data, making it difficult to identify starting from user behavior patterns, user interest in certain things and user habits. This research aims to explore scattered text data as text data has not been analyzed deeply in terms of word structure, thus concealing a lot of information and making it difficult to conduct analysis to determine user patterns and behavior on social media. The results obtained from this research are in the form of analysis and models that can identify and understand user patterns and behavior on social media using association rules and machine learning approaches. In applying the association technique, an a priori algorithm is used, which in the process determines all text data into item sets so that it can identify the habits of social media users and in the machine learning process there is a model formation process so that the results can be compared. There are several stages in the process of determining user habits, such as cleaning data, changing unstructured data to structured, then going through the rule association stage using an a priori algorithm applied to social media data so that the relationship can be seen between words. After that, machine learning is applied so that comparisons occur in seeing the relationships between words. This is new research in producing a model to identify user behavior using association rules and machine learning so that it can be used to map positive, negative or neutral user behavior