Publication

Abstract : Social networking portals serve as an ideal platform for a person or an organization, to accomplish self-presentation and self-enhancement goals there by to understand their social relevance and hence, there have been many studies attempting to identify the relationship between different aspects of social media articles. Machine learning methods play a critical role in social media data analytics. This study evaluates the efficiency of predictive analytics with social media data of a private healthcare institution using classification and clustering algorithms. In this study, we also investigated the influence of system specific feature set and user generated features of social media articles to identify relevant data mining algorithms that are suitable for identifying knowledge-related patterns in dataset. Among the classification methods it is found that Bayesian algorithm performs better compared to other classification techniques. K-Means and Filtered cluster among the clustering techniques have better predictive analytics efficiency compared to other algorithms. Identifying the predictive analytics efficacy of these algorithms helps healthcare institution to build a model with most appropriate characteristics of social media articles.