Colorectal cancer pathology text relation extraction method

• Main Authors: CAI NIAN, TIAN YINFENG, LUO ZHIHAO, LI SHISHAN, CHEN CHUANWEN, WANG HAN
• Format: Patent
• Language: Chinese; English
• Published: 21.11.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a colorectal cancer pathological text relation extraction method. The method comprises the following steps of S1, performing double-flow feature extraction on a text; s2, performing multi-task extraction based on feature decomposition on the extracted double-flow features to obtain subjects and objects and a relationship; and S3, carrying out triple alignment by using a three-affine mechanism according to the extracted subjects, objects and relationships. According to the method, double-flow feature extraction and multi-task decomposition pathological text relation extraction are used, and the triad is generated according to the three-affine score, so that the problems that the extraction accuracy of the triad is low, independence and relevance of the relation and the features between the subject and the object are not considered, and information interaction between the relation and the subject and the object is not further considered are solved. 本发明公开了一种结直肠癌病理文本关系抽取方法，包括以下步骤：S1：对文本进行双