Overview / Usage

Primarily, the dissemination of misinformation related to Covid-19 will be analyzed. The tweets will be converted into word vectors by fine-tuning this data on the word embedding generation models such as Word2Vec, GloVe, Doc2Vec, BERT, ELMO, and FastText, etc. Furthermore, user-groups that are prone to trusting such misinformation will be identified using the retweets and the user-base information. This study will help in analyzing the misinformation spread and the user-groups originating and getting affected by such information. Later, the geographical hotspots of generation and spread of this information will be identified. This will help in associating the geographical spread of coronavirus with the spread of such misinformation.

The word embedding generation models such as Word2Vec, Doc2Vec, and BERT are pre-trained and needed to be fine-tuned on the topic-related data to produce better word embeddings. However, these models require heavy hardware and processing time. Using the oneAPI provided resources such as hardware and software can help in the current work.