Medical risk information extraction based on Hidden Markov Model

As the number of explosive growth in the medical literature, how to extract from the medical literature for clinical useful risk information has always been a challenge. To address the task, the integrated model is proposed. In this paper, the proposed model mainly includes two parts. First, risk statements are extracted from unstructured medical literature text for breast cancer, focusing on probability expression. Second, the causing events, influenced events and cue words in risk statements are analyzed based on Hidden Markov Model. To a certain extent, the approach can reduce the manual workload of extracting risk information, improve the efficiency of extraction, and lay the foundation of knowledge acquisition for clinical decision support system.