A Natural Language Processing System for Truth Detection and Text Summarization

In the futuristic world, differentiating between accurate information and false information would be challenging. To identify the fake news a truth detection system will be extremely useful. Different artificial intelligence models were compared, and the highest performing algorithm was chosen. This model was subsequently updated using a larger data-set in order to achieve better results in more real-world applications. The most essential assumptions are that the user will only utilize the subject field in which the model will work. Earlier works have implemented the process of fake news and truth detection separately from the process of summarizing the said news/information. No method was available for streamlining the process to detect whether given information was true or not and to then summarizes the given information automatically. This study has compared and contrasted the differences in accuracy and overall performance of each algorithm for fake news detection and summarization of text. Using smaller dataset, LSTM cells outperformed all other models with an accuracy of 99%. For the larger dataset, the modified LSTM has worked really well with 98.6 percent accuracy.