Machine learning methods solving an inverse problem in spectroscopy: comparison of efficiency and noise resilience

Problems of indirect measurements arising in experimental science belong to the class of inverse problems (IP), which are often ill-posed, non-linear, and have high dimensionality. Machine learning methods are used to solve IP due to their ability to resist these unfavorable properties. In this study, we solve an IP of determination of concentrations of components in multi-component solutions by their Raman spectra. The results demonstrated by various machine learning methods are compared by the solution error and by their resilience to various types of noise encountered in experimental spectroscopy. The best results were demonstrated by multi-layer perceptrons with two hidden layers and by convolutional neural networks with one convolutional layer. For multiplicative noise with high noise level, the most noise resilient algorithms were random forest and gradient boosting.

[1]  François Chollet,et al.  Keras: The Python Deep Learning library , 2018 .

[2]  I. G. Persiantsev,et al.  Neural Network Approaches to Solution of the Inverse Problem of Identification and Determination of Partial Concentrations of Salts in Multi-сomponent Water Solutions , 2014, ICANN.

[3]  M. Fontana,et al.  Inorganic salts diluted in water probed by Raman spectrometry: Data processing and performance evaluation , 2015 .

[4]  Sergey Burikov,et al.  Joint Application of Group Determination of Parameters and of Training with Noise Addition to Improve the Resilience of the Neural Network Solution of the Inverse Problem in Spectroscopy to Noise in Data , 2018, ICANN.

[5]  N. Gros,et al.  Expert system for the ion chromatographic determination of alkali and alkaline earth metals in mineral waters , 1995 .

[6]  J. Saja,et al.  Effect of electrolyte concentration on the Raman spectra of water in aqueous solutions , 1986 .

[7]  M.H. Hassoun,et al.  Fundamentals of Artificial Neural Networks , 1996, Proceedings of the IEEE.

[8]  D. Pletcher,et al.  Amperometric method for the determination of nitrate in water , 1995 .

[9]  G. Raber,et al.  Voltammetric determination of nitrate with a modified carbon paste electrode , 1994 .

[10]  B. Saad,et al.  Analysis of anions and cations in drinking water samples by Capillary Ion Analysis , 1998 .

[11]  Alexander LeNail,et al.  NN-SVG: Publication-Ready Neural Network Architecture Schematics , 2019, J. Open Source Softw..

[12]  Celio Pasquini,et al.  Automated gravimetric management of solutions. Part 2. Automated gravimetric approach to direct potentiometry and kappa number determination , 1995 .

[13]  A. Efitorov,et al.  Use of Adaptive Methods to Solve the Inverse Problem of Determination of Composition of Multi-Component Solutions , 2018, Optical Memory and Neural Networks.

[14]  Xiangfeng Liang,et al.  Simultaneous detection of anions and cations in mineral water by two dimensional ion chromatography. , 2018, Journal of chromatography. A.

[15]  Alexander Efitorov,et al.  Study of Efficiency of Dividing the Problem Space as a Means to Improve Solution of Multi-parameter Inverse Problem by Adaptive Methods , 2018 .

[16]  Alexander Efitorov,et al.  Significant Feature Selection in Neural Network Solution of an Inverse Problem in Spectroscopy , 2015 .

[17]  Simon Haykin,et al.  Neural Networks and Learning Machines , 2010 .

[18]  T. Crompton.  Determination of Anions in Natural and Treated Waters , 2002 .

[19]  R. Singh,et al.  Comparative Evaluation of Trace Heavy Metal Ions in Water Sample using Complexes of Dithioligands by Flame Atomic Absorption Spectrometry , 2018 .

[20]  B. Hart,et al.  Determination of metal ion mixtures using pH gradient flow-injection analysis with fluorescence detection , 1995 .

[21]  S. Burikov,et al.  The effect of hydration of ions of inorganic salts on the shape of the Raman stretching band of water , 2005 .

[22]  Tatiana A. Dolenko,et al.  Valence band of liquid water Raman scattering: some peculiarities and applications in the diagnostics of water media , 2000 .

[23]  Weidong Yang,et al.  Use of Artificial Neural Networks in Near-Infrared Spectroscopy Calibrations for Predicting Glucose Concentration in Urine , 2008, ICIC.