“Ensemble Incremental Model” for Misinformation Detection in Healthcare

With the growth of social media usage, information sharing has become a concern due to the spread of misinformation and its strong impact on society. Researchers have developed innovative techniques to detect and classify false information online. Improving the accuracy of these models using ensemble learning methods has gained popularity recently. However, these ensemble models are based on standard machine learning and ensemble classifiers and don't consider the novel classifiers related to false information detection. In addition, the ensemble models developed so far consider only textual features, leading to over fitting the models. Further, the changing data is a concern as the models show poor performance with old data. Considering these limitations, authors have proposed the 'Ensemble Incremental Model (ElM),’ which extracts novel user and textual features, implements the ensemble method using recently developed Content Similarity Measure (CSM) algorithm and standard machine learning classifiers, and incorporates incremental learning approach. It was observed that the novel approach EIM showed the highest accuracy of 93.44%.