Medicine data

Machine learning offers a wide range of techniques to predict medicine expenditures using historical expenditures data as well as other healthcare variables. For example, researchers have developed multi-layer perceptron (MLP), long-short term memory (LSTM), and convolutional neural networks (CNN) for predicting healthcare outcomes. However, recently proposed generative approaches (e.g., generative adversarial networks; GANs) are yet to be explored for time-series prediction of medicine-related expenditures. The primary objective of this research is to develop and test a generative adversarial network model (called “variance-based GAN or V-GAN”) that specifically minimizes the difference in variance between model and actual data during model training. For our model development, we used patient expenditure data of a popular pain medication in the US. In the V-GAN model, we used an LSTM model as a generator network and a CNN model as a discriminator network. The performance of V-GAN model was compared with other GAN variants and a baseline LSTM model. Results revealed that the V-GAN model outperformed other GAN-based prediction models and the LSTM model in correctly predicting future medicine expenditures of patients. Through this research, we highlight the utility of using GAN-based architectures involving variance minimization for predicting patient-related expenditures in the healthcare domain.

机器学习为利用历史支出数据及其他医疗变量预测医药支出提供了丰富的技术手段。例如，研究者们已开发出多层感知器（MLP）、长短期记忆网络（LSTM）以及卷积神经网络（CNN）等模型来预测医疗结果。然而，近期提出的生成式方法（例如，生成对抗网络；GANs）在预测与医药相关的时序支出方面尚未得到充分探索。本研究的首要目标是开发并测试一种生成对抗网络模型（命名为“基于方差的最小化GAN或V-GAN”），该模型在模型训练过程中专门最小化模型与实际数据之间的方差差异。在模型开发过程中，我们采用了美国一款流行止痛药的病人支出数据。在V-GAN模型中，我们使用了LSTM模型作为生成网络，CNN模型作为判别网络。V-GAN模型的表现与其他GAN变体及基准LSTM模型进行了比较。结果显示，V-GAN模型在正确预测患者未来医药支出方面优于其他基于GAN的预测模型以及LSTM模型。通过本研究，我们强调了在医疗领域预测与患者相关的支出时，使用基于GAN架构并涉及方差最小化的方法的价值。