Disclosed is a method for security research report analysis using deep learning models. The method comprises: receiving a security research report to be analysed (S01); carrying out sentence segmentation on the security research report to obtain each sentence (S02); scoring each sentence by using a first deep learning model to obtain a first score of each sentence, and determining a viewpoint-type sentence based on the first score (S03); scoring each viewpoint-type sentence by using a second deep learning model to obtain a second score of each viewpoint-type sentence, and determining an industry viewpoint-type sentence based on the second score (S04); scoring each industry viewpoint-type sentence by using a third deep learning model to obtain a third score of each industry viewpoint-type sentence (S05); and determining an overall emotion trend of the security research report according to the second and third scores of each industry viewpoint-type sentence (S06). The analysis of a security research report is realized by means of intelligent scoring using deep learning models, the problem of the efficiency and accuracy of an analysis scheme for the security research report being relatively low can be solved, and the efficiency and accuracy of the analysis of the security research report are improved.
展开▼