数据挖掘在艾滋病病人CD_~4+T淋巴细胞与机会性感染关系的应用研究

目的利用数据挖掘技术分析艾滋病(AIDS)病人CD4+T淋巴细胞与机会性感染的关系,以期对合并机会性感染的AIDS病人的早期预防性用药提供决策支持。方法使用数据挖掘中的C4.5决策树算法,分析重庆市公共卫生医疗救治中心2003-2008年的207例AIDS病人的相关数据。结果 AIDS病人的CD4+T淋巴细胞值在A(0-50)区间,合并感染卡氏肺孢子虫肺炎(PCP)机会很大(概率为82.35%)。如果没有感染PCP,但是感染了隐球性脑膜炎(NMY),其CD4+T淋巴细胞值在A(0-50)区间;既没有感染PCP,也没有感染NMY,但合并感染丙型肝炎(丙肝)及乙型肝炎(乙肝)者,其CD4+T淋巴细胞值在D(201-300)区间;没有感染PCP,也没有感染NMY,但合并感染丙肝未感染乙肝,其CD4+T淋巴细胞值在C(101-200)区间。结论通过使用数据挖掘技术中的决策树算法,得出AIDS病人CD4+T淋巴细胞值在某一个区间,更容易合并某一种机会性感染,CD4+T淋巴细胞值与机会性感染有着重要的关系。

Exploratory development of the data mining in the relationship between CD4+ cell counts and the opportunistic infection among AIDS patients

Objective To analyze the relationship between CD+4 cell counts of AIDS patients and the opportunistic infection by utilizing the data mining technology,and to provide decision-making support for an early preventive medication among AIDS patients who were complicated with opportunistic infections.Methods The C4.5 decision tree algorithm of data mining was considered as the main method.Interrelated data of 207 AIDS patients from Chongqing Public Health and Medical Service Center between 2005 and 2008 were collected as study samples.Results AIDS patients whose CD+4 cell value lay in A(0-50) sector had a higher opportunity to be infected with pneumocystis carinii pneumoni(PCP);for those infected with cryptococcal meningitis(NMY)without PCP,the CD+4 cell value lay in A sector.For those who were co-infected with hepatitis C and hepatitis B,but with neither PCP nor meningitis,the CD+4 cell value lay in D(201-300)sector;if the patients were only co-infected with hepatitis C,their CD+4 cell value lay in C(101-200)sector.Conclusion CD+4 counts in some sectors have close relationship with the occurrence of some types of opportunistic infection.