RAPOT: A Fine-Grained Risk Assessment Framework on Public Opinion for Trial Management

Abstract

Since public opinion from social media has a growing impact and supervision on the trial, risk assessment on public opinion is increasingly important in the refinement of trial management. However, due to insufficient historical training data, lack of high accuracy in risk prediction and so on, existing methods cannot be effectively applied to risk assessment on public opinion. To address this, we propose a fine-grained risk assessment framework on public opinion with fuzzy numbers. Firstly, we establish a multilayer indicator model for assessing the Risk of Public Opinion (POR), in the model, the indicators are defined and allocated priorities based on AHP. Secondly, we design a public opinion analysis module for indicator evaluation, which includes analysis in public opinion sentiment, in hot search and in social media coverage. Specially, the public opinion sentiment is calculated by aggregating sentiments of topic clusters, which is accurate and reasonable. Finally, the fuzzy number similarity is employed to determine the level of POR in the nine-levels risk system. Experimental results validate the efficiency of our framework when assessing the POR.