Artificial Neural Network and Savitzky Golay Derivative in Predicting Blood Hemoglobin Using Near-Infrared Spectrum

Abstract

Monitoring blood hemoglobin level is essential to diagnose anaemia disease. This study aims to evaluate the capability of an artificial neural network (ANN) and Savitzky Golay (SG) pre-processing in predicting the blood hemoglobin level based on the near-infrared spectrum. The effects of the hidden neuron number and different SG pre-processing strategies were examined and discussed. ANN coupled with first order SG derivative and five hidden neurons achieved better prediction performance with root mean square error of prediction of 0.3517 g/dL and coefficient determination of prediction of 0.9849 compared to the previous studies. Results depict that ANN that coupled with first order SG derivative could improve near-infrared spectroscopic analysis in predicting blood hemoglobin level, and the proposed nonlinear model outperforms linear models without variable selections. This finding suggests that the modelling strategy is promising in establishing a better relationship between the blood hemoglobin and near-infrared spectral data.

A Bootstrapping Soft Shrinkage Approach and Interval Random Variables Selection Hybrid Model for Variable Selection in Near-Infrared Spectroscopy

“…Using the IDRC shootout 2010 software, Karl Norris [38] has produced this dataset that has been used by Mohd Nazrul Idrus [39]. With the spectrometer of NIR Systems 6500, the blood samples have been analyzed.…”