Deep Learning Estimation of Medical Substance Concentrations Using Pytorch

Machine learning has been widely used in healthcare services. Accurate estimation of target analyte concentrations is crucial for disease diagnosis and treatment planning. In this work, medical datasets are collected and pre-processed. A deep neural network model is constructed based on Pytorch. By evaluating the model and algorithm performances, the application potential of accurate biomarker identification is revealed for disease diagnosis. The estimation is tested to be reliable in discriminating the type and starting concentration of the drug based on the kinetic characteristics over time.