Integrated Classification and Regression for Quantitative Biomarker Analysis in Mixture Samples

This study aims to develop a comprehensive approach for the classification and concentration estimation of multiple target analytes in biosensing samples by combining classification and regression algorithms. This paper presents an algorithm called “post-classification regression”, which divides the data into different categories and applies the corresponding regression model to each category to estimation the concentrations. In experiments, we first used an advanced classifier to classify the data and then linked each data point to a regression model of its category. In this way, we can make full use of the advantages of the classification model and regression model to achieve accurate classification of multiple target analytes and accurate concentration estimations.