基于R的Apriori算法在超长住院患者信息挖掘中的应用

The Application of Apriori Algorithm based on R Language in Data Mining of Long Hospitalized Patients

目的 使用Apriori算法查找某院2014年-2015年住院天数超过30天的超长住院患者病案首页各指标中的关联规则,以期能够分析超长住院的内在原因,并且为缩短患者的住院天数提供思路。方法 利用R语言的arules包中的生成关联规则的apriori函数,编写程序将数据导入系统,探索选取的2011版病案首页中超长住院日患者的包括性别、年龄、出院诊断、是否手术等20个指标是否存在关联规则,并分析其原因。结果 根据编写的程序,共获得329 834条强关联规则,得到如下的规则住院天数为31天~40天且出院诊断首位为Z的患者,其费用一般为50000元以下;呼吸病区的患者住院期间一般需使用抗菌药物;骨科出院患者的年龄多为19岁~30岁;神内病区、呼吸病区、肿瘤病区出院的超长住院患者通常不进行手术治疗;而心外病区、骨科病区、普外病区则通常要进行手术治疗。结论 通过关联规则分析,可以找到超长住院的原因,为减少患者的住院天数提供思路。

objective To use Apriori algorithm to find a hospital in 2014 and 2015 hospital stay more than 30 days long hospitalization patients in the firstpage of medical records of each index in the association rules, in order to analyze the internal reason overstay hospitalization, and shorten the patient＇s hospital stay to provide ideas. Methods Using apriori function in package rules of R language, then program the code segment and include data into the R system including gender, age, discharge diagnosis, and also others, whether there is association rules, and analyze the reasons. Results According to the procedures for the preparation were obtained 329834 strong association rules, follows the rules for the duration of hospitalization between 31 and 40 days, and the first discharge diagnosis started with Z, the cost is generally below 50000 yuan; respiratory ward patients hospitalized during the general use of antibiotics; orthopaedic hospital patient＇s age is 19-30 years old; God ward, respiratory ward, tumor ward discharged long hospitalized patients do not usually carry out surgical treatment; and heart disease, orthopedic ward, general surgery ward is surgery. Conclusion Through the analysis of association rules, we can find the reasons for the length of stay in hospital, and provide ideas for reducing the number of days of hospitalization.