摘要
在医疗领域数字化转型进程中,传统数据分析方法已无法处理庞大、复杂且异构的医疗数据,因此研究提出了一种融合改进Apriori算法与反向传播神经网络的医疗大数据挖掘与疾病预测模型。研究使用改进Apriori算法医疗大数据挖掘,并在此基础上利用反向传播神经网络预测疾病,最后对研究模型进行测试。测试结果显示,在0.05~0.20支持度区间,研究模型的运行时间最快,当支持度为0.20时,运行时间仅0.1 s,此时决策树模型、支持向量机模型、逻辑回归模型的运行时间分别为0.19、0.16、0.12 s。训练数据预测的混淆矩阵显示,模型对患病与未患病样本的正确预测数分别达22409与21437,模型错误预测为患病的未患病样本数7591,模型错误预测为未患病的患病样本数8563。由研究结果可知,此次研究能够为医疗大数据深度分析提供新路径,对推动医疗模式向精准预防与个性化治疗转型具有重要意义。
In the process of digital transformation in the medical field,traditional data analysis methods have been unable to handle huge,complex and heterogeneous medical data.Therefore,a medical big data mining and disease prediction model integrating the improved Apriori algorithm and backpropagation neural network is proposed.The study uses the improved Apriori algorithm for medical big data mining,and on this basis,utilizes the backpropagation neural network to predict diseases.Finally,the research model is tested.The test results show that within the support degree range of 0.05-0.20,the running time of the research model is the fastest.When the support degree is 0.20,the running time is only 0.1 s.At this time,the running times of the decision tree model,the support vector machine model,and the logistic regression model are 0.19 s,0.16 s,and 0.12 s respectively.The confusion matrix predicted by the training data shows that the correct predictions of the model for the diseased and non-diseased samples reach 22409 and 2437 respectively.The number of non-diseased samples wrongly predicted by the model is 7591,and the number of diseased samples wrongly predicted by the model is 8563.The research results show that this study can provide a new path for in-depth analysis of medical big data and is of great significance for promoting the transformation of the medical model towards precise prevention and personalized treatment.
作者
俞利张
YU Lizhang(Shaoxing Second Hospital Medical Community General Hospital,Shaoxing 312000,China)
出处
《国外电子测量技术》
2025年第11期257-262,共6页
Foreign Electronic Measurement Technology
基金
医院云计算平台业务资源优化与安全协同探索研究(2024XHYS-Z03)。
