WSPXY Combined with BP-ANN Method for Hemoglobin Determination Based on Near-Infrared Spectroscopy

Han Tian,Linna Zhang,Ming Li,Yue Wang,Dinggao Sheng,Jun Liu,Chengmin Wang
DOI: https://doi.org/10.1016/j.infrared.2019.103003
IF: 2.997
2019-01-01
Infrared Physics & Technology
Abstract:Hemoglobin concentration is commonly used in clinical medicine to diagnose anemia, and manage red blood cell transfusions. Spectral methods were advantageous at fast and non-reagent measurement. An appropriate method for calibration set selection was very important for a robust quantitative model. In this paper, Weighted sample set partitioning based on joint x-y distances (WSPXY) method was applied for sample selection in calibration set. Moreover, proper stoichiometry method can be greatly beneficial for improving the prediction performance of quantification model. Partial Least Square Regression (PLSR) method was commonly used in composition analysis with spectroscopy methods. Back Propagation Artificial Neural Network (BP-ANN) was proved to be effective for quantification analysis. In this research, BP-ANN method was used for prediction of haemoglobin concentration. Compared with the results from PLSR method, WSPXY combined BP-ANN method could enhance the prediction performance of haemoglobin content. The overall results convinced that BP-ANN method was a reliable and robust method for haemoglobin determination. This research can provide valuable references for hemoglobin concentration determination by spectral methods.
What problem does this paper attempt to address?