基于关联规则和人工神经网络的大肠癌中医证治规律研究
本文选题:关联规则 + 人工神经网络 ; 参考:《广州中医药大学》2017年硕士论文
【摘要】:目的:通过对当代中医治疗大肠癌的相关医案文献进行整理和量化研究,以期探讨中医诊治大肠癌的临床规律和建立大肠癌证候预测模型,为临床辨证治疗大肠癌提供规范和参考。方法:1.通过全面检索1979-2015年各大论文数据库中收录有关中医药治疗大肠癌的文献,建立当代中医辨证治疗大肠癌的文献数据库。对使用中药、四诊信息、证候等进行频数分析,对中药与中药、四诊信息与中药、证候与中药进行关联分析。2.将收集到的大肠癌医案分解为四诊信息和证候模块录入Excel软件中,并导入到IBM SPSS 19.0中,把"脾肾阳虚证"设置为因变量,把收集到四诊信息作为因子,用其中70%的数据进行建模训练,剩余30%的数据作为检验,运用多层感知器和径向基函数两种算法进行大肠癌证候预测模型研究。结果:1.在137篇大肠癌医案中共使用中药203味,用药总频次1760次,常用中药(用药频数15%)有24味。高频中药有白术、黄芪、茯苓、薏苡仁、白花蛇舌草、党参、陈皮等,中药主要集中分布在补气药、清热药、理气药、利水消肿药、化湿药、清热解毒药、清热燥湿药、补血药。医案共收录大肠癌伴随症状70种,共699频次。大肠癌常见症状有腹痛、乏力、腹胀、便血、体瘦等主要伴随症状;苔薄白、舌淡、苔白、脉细、舌质红为常见舌脉象。大肠癌篇文献中,筛选出证型有29个,经过归类整理后得到8个证型。证候分型以气血亏虚较多,其次是湿热内蕴、气滞血瘀、淤毒内阻、脾失健运。2.挖掘出中药与中药的关联度较高的是:茯苓分别与白术、陈皮;木香分别与茯苓、白术;党参分别与茯苓、白术;甘草分别与黄芪、白术;陈皮分别与黄芪、白术;白术与黄芪。四诊与中药的关联统计结果:脉细、便溏、苔薄白、口苦、舌质红、口淡的大肠癌患者采用黄芪、白术等补气药较多。证候与中药关联结果:与脾肾阳虚密切相关的有黄芪、白术、陈皮、山药和茯苓;与脾失健运密切相关的有党参、茯苓、白术;与脾虚湿困密切相关的有白芍、茯苓、陈皮和木香;与气血亏虚相关密切相关的有黄芪、白术和薏苡仁;与气滞血瘀密切相关的有黄芪、茯苓和白术;与湿热内蕴密切相关的有甘草、薏苡仁、黄芪。3.基于多层感知器的大肠癌脾肾阳虚证候模型的训练组的样本有92例,占67.2%;测试组的样本有45例,占32.8%。样本训练组的正确百分比为95.7%,测试组的正确百分比为93.3%,ROC曲线范围值为0.972。基于径向基函数的大肠癌脾肾阳虚证候模型的训练组的样本有102例,占74.5%;测试组的样本有35例,占25.5%。样本训练组的正确百分比为90.2%,测试组的正确百分比为82.9%。ROC曲线范围值为0.760。基于多层感知器的大肠癌脾肾阳虚证候预测模型计算出四诊信息重要性前六位依次为苔薄脉沉细纳呆腹胀脉细睡眠差,基于径向基函数的大肠癌脾肾阳虚证候预测模型计算出四诊信息重要性前六位依次为苔薄白腹胀舌淡体瘦纳呆腹痛。结论:从用药频数分析,当代中医医案处方中采用白术、黄芪、茯苓等24味药物较多,其中以补气药、理气药、清热药和化湿药较多;从四诊信息上看,腹痛、乏力、腹胀、便血等为大肠癌患者的主要伴随症状,证候频数显示大肠癌多以湿热内蕴、气血亏虚为主。本研究中药与中药关联分析研究结果显示,置信度和支持度最高的药对主要为补血药与补气药形成组合药对,理气药与补血药组成药对。从四诊信息与中药的关联分析可知,脉细、便溏、苔薄白的大肠癌患者采用黄芪、白术、茯苓等补气药、利水消肿药较多。从证候与中药关联分析可知,脾肾阳虚、脾失健运、脾虚湿困的患者应多用白术、黄芪、茯苓等益气健脾的中药。基于多层感知器的大肠癌脾肾阳虚证候预测模型ROC曲线范围值较为接近1,说明此预测模型准确性较高。可优先考虑选用基于多层感知器的预测模型研究大肠癌证候的判别。
[Abstract]:Objective: to organize and quantify the relevant medical records of modern Chinese medicine for colorectal cancer, in order to explore the clinical rules of the diagnosis and treatment of large intestine cancer by traditional Chinese medicine and to establish the model of the prediction of the syndrome of large intestine cancer, and to provide a standard and reference for the treatment of colorectal cancer in clinical syndrome differentiation. 1. Literature on the treatment of colorectal cancer by traditional Chinese medicine and the establishment of a literature database for the treatment of large intestine cancer by contemporary TCM syndrome differentiation, frequency analysis on the use of traditional Chinese medicine, four diagnosis information and syndrome, and the correlation analysis of traditional Chinese medicine and traditional Chinese medicine, four diagnosis information and traditional Chinese medicine, syndrome and traditional Chinese medicine by.2., the medical cases of large intestine cancer are decomposed into four diagnosis information and syndrome modules In the Excel software, and imported into the IBM SPSS 19, the "spleen and kidney yang deficiency syndrome" was set as the dependent variable, and the four diagnosis information was collected as a factor, and 70% of the data were used to model the modeling training. The remaining 30% data were used as the test, and the multi-layer perceptron and radial basis function were used to study the prediction model of the syndrome of large intestine cancer. The results were as follows: 1. In 137 large bowel cancer cases, 203 flavors of traditional Chinese medicine were used, the total frequency of medicine was 1760 times, and the common Chinese medicine (frequency 15%) had 24 flavors. High frequency traditional Chinese medicine had Atractylodes macrocephala, astragalus, Poria, coix seed, white flower Hedyotis, Codonopsis, tangerine and so on. The Chinese traditional medicine was mainly concentrated in the medicine of Tonifying Qi, the antipyretic drugs, the antipyretic drugs, the antipyretic drugs, the antipyretic drugs, and the heat clearing dryness. A total of 70 types of colorectal cancer accompanied by symptoms, including abdominal pain, fatigue, abdominal distention, blood, thin body and other main symptoms, the common symptoms of large bowel cancer, such as thin white, pale tongue, whiteness, vein fine and tongue red as common tongue pulse. In the literature of large intestine cancer, 29 types were screened, and 8 syndromes were obtained after classification. There are more symptoms of Qi and blood deficiency, followed by damp heat accumulation, qi stagnation and blood stasis, silt resistance, spleen loss and transport of Chinese traditional medicine and traditional Chinese medicine have higher correlation degree: tuckahoe and Baizhu, tuckahoe, Atractylodes macrocephala respectively with Poria cocos and Atractylodes macrocephala; Radix Astragalus, Atractylodes rhizome, rhizoma Atractylodes, Rhizoma Atractylodes, Rhizoma Atractylodes; Rhizoma Atractylodes; Rhizoma Atractylodes; and Rhizoma Atractylodes, respectively, with Radix Astragalus, Atractylodes and.2. Astragalus. Four diagnosis and traditional Chinese medicine correlation statistical results: pulse thin, loose stool, thin white moss, oral pain, tongue quality red, and tonifying large intestine cancer patients with Astragalus, Atractylodes and other drugs. Closely related to the deficiency of Qi and blood, including Astragalus, Rhizoma Atractylodes and coix seed, closely related to deficiency of Qi and blood deficiency, which closely related to qi stagnation and blood stasis are Astragalus, Poria and Atractylodes, and there are Glycyrrhiza, coix seed and Huangqi.3., which are closely related to the damp and blood stasis, and the syndrome model of spleen and kidney yang deficiency syndrome of large intestine cancer based on multilayer perceptron The sample of the training group was 92, accounting for 67.2%; the sample of the test group had 45 cases, the correct percentage of the 32.8%. sample training group was 95.7%, the correct percentage of the test group was 93.3%, the ROC curve range was 0.972. based on the radial basis function of the spleen kidney yang deficiency syndrome model in 102 cases, accounting for 74.5%, and the sample of the test group. 35 cases, the correct percentage of the 25.5%. sample training group is 90.2%, the correct percentage of the test group is the 82.9%.ROC curve range value is 0.760. based on the multi-layer perceptron based on the spleen and kidney yang deficiency syndrome prediction model of the colorectal cancer. The first six places of the four diagnosis information are the thin dyke sinks and the abdominal distention pulse fine sleep poor, based on the radial basis function The prediction model of spleen and kidney yang deficiency syndrome of large intestine cancer calculated the importance of the first six cases of the four diagnosis information. Conclusion: from the analysis of the frequency of medicine, there are 24 kinds of drugs in the prescription of modern medicine, such as Atractylodes, astragalus, Poria, and so on. Information on information, abdominal pain, fatigue, abdominal distention, blood and so on as the main concomitant symptoms of large intestine cancer patients, the frequency of syndrome shows that large intestine cancer is more and more damp and hot, Qi and blood deficiency is the main. According to the correlation analysis between the four diagnosis information and the Chinese medicine, we can see that the patients with large intestine cancer with pulse thin, loose stool and thin white fur are filled with Astragalus, Atractylodes and poria, and there are many drugs for eliminating swelling. The ROC curve range of the prediction model of spleen and kidney yang deficiency syndrome based on multilayer perceptron is close to 1, which indicates that the prediction model is more accurate and the prediction model based on multilayer perceptron can be preferred to study the identification of the syndrome of colorectal cancer.
【学位授予单位】:广州中医药大学
【学位级别】:硕士
【学位授予年份】:2017
【分类号】:R273
【参考文献】
相关期刊论文 前10条
1 李施新;李远;刘孟敏;;基于关联规则和复杂系统熵聚类的100位全国名老中医治疗冠心病证治规律研究[J];中华中医药杂志;2016年04期
2 李悠然;谷云飞;陈邑岐;王浩;;四君子汤加减联合化疗对结直肠癌患者的Meta分析[J];中国实验方剂学杂志;2016年06期
3 郑树;张苏展;黄彦钦;;结直肠癌研究30年回顾和现状[J];实用肿瘤杂志;2016年01期
4 孙玉姣;李盼盼;原嘉民;徐福平;黄鹂;杨志敏;;国医大师颜德馨含川芎膏方用药规律的数据挖掘[J];时珍国医国药;2015年12期
5 熊江华;张昭;倪寅;李艳;;基于数据挖掘探讨李济仁教授治疗慢性胃炎的用药规律[J];中医药临床杂志;2015年11期
6 李远;刘盼英;李海滨;李涛;;运用数据处理技术探讨原发性血小板增多症的中医辨证和用药规律[J];医学研究与教育;2015年05期
7 吴皓萌;徐志伟;敖海清;;基于数据挖掘的针灸治疗慢性泄泻的选穴规律研究[J];中华中医药杂志;2015年12期
8 刘宣;柴妮;韩植芬;刘慧;邓皖利;周利红;隋华;季青;李琦;;健脾解毒方对湿热证结肠癌小鼠肿瘤血管新生的抑制作用[J];上海中医药大学学报;2015年06期
9 黄坤艳;程军平;张世忠;张定棋;唐翠娟;喻欢;谭凌菁;梅志刚;;基于数据挖掘的糖尿病胃轻瘫中医用药规律研究[J];时珍国医国药;2015年09期
10 郭立忠;;补益药党参的药理作用与临床应用研究[J];中国卫生标准管理;2015年22期
,本文编号:1959631
本文链接:https://www.wllwen.com/zhongyixuelunwen/1959631.html