Optimizing the Influence of Temporal Dynamics, Network Topologies, and Semantics on Unsupervised NLP Algorithms

Abstract- The purpose of this study was to generate an algorithm able to decipher bots in social media. Prior research shows that variables/parameters affect the detection of AI; however, none attempt to compile an algorithm accurate enough to be deployed into a real-world scenario. Data was collected through mixed methods, in which data was collected online and through questionnaires. Participants included individuals from all demographics, only restricted to demonstrate no bias. Initial results show a strong correlation with variables on the usage of AI. This means that a model which can effectively deduce the usage of AI is plausible. Therefore, the conclusion can be made that it is possible to find bots in social media; however, this is limited to 70% accuracy, given the available resources. Future research should be targeted towards making sure text can be deciphered with more accuracy.