Method, device and system for extracting symptoms and causes in medical text

The invention discloses a method, a device and a system for extracting symptoms and causes in a medical text, and particularly relates to the technical field of natural language processing information extraction. The system mainly comprises a reading module, an extraction module, a sentence pattern module and a display module. The reading module is mainly used for reading some medical texts by the system. The extraction module mainly comprises a relational word extraction unit, a symptom extraction unit and a symptom cause extraction unit. The method mainly comprises the following steps: firstly, summarizing a partial relational word set; constructing sentence patterns through the symptoms and the relational words to extract symptom causes; and then performing extracting through the entity subjected to impurity removal and updating; and finally, evaluating an extraction result, if the number of the extracted entities is increased compared with that of the existing entities, continuing iteration, and otherwise