论文部分内容阅读
卵巢浆液性肿瘤分成良性、交界性和恶性三类,因病理医师在诊断时常有差异.所以我们应用模糊数学对286例(312个瘤体)进行客观分类.作为研究的基础与方法鉴定的标准。然后得到各级瘤体的模式样本与隶属函数。用模糊模式识别对这些病例重新诊断。结果,符合率为99.36％(310/312),对照方法:数据用微机获得多元回归方程后用方程重新诊断,结果符合率为95.83％(299/312)。讨论分析了该肿瘤分类困难的原因。 为日常工作方便,制作了微机诊断软盘,能按日常病理工作要求很快打印出正规病理报告书。
Ovarian serous tumors are divided into three categories, benign, borderline and malignant, due to pathologists often differ in the diagnosis, so we use fuzzy mathematics to objectively classify 286 cases (312 tumors) as the basis for the study and method identification standards . Then we get the model samples and membership functions of the tumor at all levels. Re-diagnosis of these cases with fuzzy pattern recognition. The results showed that the coincidence rate was 99.36% (310/312). The control method: The data were re-diagnosed by the equation after obtaining the multiple regression equation by computer, and the coincidence rate was 95.83% (299/312). Discuss the reasons for the difficulty of the classification of the tumor. For routine work convenience, made a computer diagnostic floppy disk, according to daily pathological work requirements quickly print out the regular pathology report.