Measurement study of human blood pH based on optical technique by back propagation artificial neural network

Abstract

To verification of concept, a spectroscopic method for measurement of pH in human blood through the syringe based on backpropagation artificial neural network (BP-ANN). In this paper the feasibility of design and fabricate measurement of pH was consist of 5LEDs as light source, 2 photodiodes as sensor to measure the light intensity and calculate the blood pH. The spectral data of 48 subjects were measured. Principal component analysis (PCA) was applied to deduct the dimensional of collected spectral data to reduce the infestation of redundant data. In such cases, the principal component analysis has taken as inputs of BP-ANN to correlate and predict blood pH. The calculated blood pH by BP-ANN with PCA is quite a desirable with standard error of 0.015 and 0.023 919 in validation and testing, correlations coefficient (R) 0.992 and 0.919 in validation and testing. Inspecting the accuracy of BP-ANN model results produce by statistical analysis with a relative analytical error all under 3% in validation, and testing. The results are proved that a good correlation between absorbance data with actual pH, and the model is in good agreement. Hence, the method of BP-ANN with PCA is a potential for the absorbance detection of pH in human blood through the syringe.