Study of the statistics model to be predictable of disease risk by using large-scale medical information data base

本研究的目的是利用大规模的医学信息数据，建立疾病转移预测模型。通过使用高知大学医院的医疗信息数据库中的检查值时间序列数据，构建了可预测每个患者的检查值转变的潜在曲线模型。在潜在曲线模型中，可以在评估个体数据的同时，在群体遵循它的分布背景下对其进行分析。潜在曲线模型作为能够预测个体患者数据的模型进行了研究。利用数值模拟得到的理想模拟数据，明确了模型的调整范围。

The purpose of this research is construction of the condition transition forecasting model that uses the large-scale medical information data. The latent curve model to be predictable of transition of the inspection value in each patient was constructed by using the inspection value time series data in the medical information data base at Kochi University hospital. In the latent curve model, it is possible to analyze it while evaluating individual data in the background of distribution that the group follows it. The potential curve model was examined as a model by whom the individual patient data was able to be forecast. The range of the adjustment of the model was clarified by using the ideal simulated data that had been obtained by the numerical simulation.