摘要
针对海量的医疗信息特征推送存在推送准确率低的问题,提出基于数据特征矩阵的海量医疗信息特征推送方法,采用医疗大数据特征智能采集方法获取医疗数据特征矩阵列,利用数据特征矩阵列匹配全部患者组信息和患者部分信息,向相仿度最高的患者组融入对应的患者,匹配相仿患者组里的关键词和医疗数据特征矩阵列里拟送的基础关键数据特征,患者所在的相仿患者组根据基于位置服务(LBS)优先推送方式推送医疗消息。拟送患者的未来病情特征关键词组根据患者的病情特征推算,按照LBS优先推送方式对患者推送定制医疗消息特征。实验结果表明,所提方法收集海量医疗信息特征的平均时间是4s,平均采集误差是0. 2%,进行推送测试时最高使用度和召回率分别是96. 5%和34. 5%,说明所提方法推送性能好。
In order to solve the low accuracy of mass medical information feature pushing, it proposes a mass medical information feature push method based on data feature matrix. Using medical big data feature intelligent collection method, it obtains medical data feature matrix array. It applies the data characteristic moment array to match all the patient group information and part of the patient information, and incorporates the patients with the highest degree of imitation into the corresponding patient group. Matching the keywords in the similar patient group with the basic key data characteristics to be sent in the array of medical data features, the similar patient group pushes the medical message according to the location-based service (LBS) priority push mode. Future disease to be sent to patient is according to the characteristics of the patient's illness, the key phrase of emotional feature is used to push the custom medical message characteristics of the patients according to the LBS priority push mode. The experimental results show that the average time of collecting massive medical information features is 4s and the average acquisition error is 0.2%. The maximum usage and recall rate of the proposed method are 96.5% and 34.5%, respectively. The test of proposed method has good push performance.
作者
蒋科
Jiang Ke(Nanchong Central Hospital, Sichuan Nanchong, 637000, China)
出处
《机械设计与制造工程》
2019年第3期59-63,共5页
Machine Design and Manufacturing Engineering
关键词
大数据
海量医疗信息特征
推送方法
精确采集
占用率
波动环境
large data
massive medical information characteristics
push method
accurate collection
occupancy rate
fluctuating environment