Exploratory, Regression, and Neural Network Analysis of the Stability of Cation Coronates in Selected Pure Solvents
- PDF / 868,534 Bytes
- 15 Pages / 612 x 792 pts (letter) Page_size
- 10 Downloads / 174 Views
oratory, Regression, and Neural Network Analysis of the Stability of Cation Coronates in Selected Pure Solvents N. V. Bondareva,* a V.N.
Karazin Kharkiv National University, Kharkiv, 61022 Ukraine * e-mail: [email protected]
Received May 13, 2020; revised July 29, 2020; accepted August 9, 2020
Abstract—Exploratory, regression, and neural network analysis of the stability constants of crown ether [12C4, 16C5, (CH3)216C5, DB21C7, DB24C8, DCH24C8, DB30C10] 1 : 1 complexes with alkaline (Li+, Na+, K+, Cs+, Rb+), alkaline-earth (Ca2+, Sr2+, Ba2+), and heavy (Ag+, Tl+, Co2+, Cu2+, Pb2+) metals and NH4+ in water and organic solvents (methanol, acetonitrile, acetone, N,N-dimethylformamide, nitrobenzene, nitromethane, 1,2-dichloroethane, propylene carbonate) at 298.15 K obtained via conductometry has been performed. Factor, cluster, discriminant, canonical, decision tree, regression, and neural network models of clustering, approximation, and prediction of thermodynamic constants of the complexation depending on the properties of the ligand, the cation, and the solvent have been developed. The trained MLP 7-5-5 Multilayer Perceptron Cluster has completely confirmed the k-means clustering. Independent data on the stability constants of coronates have demonstrated the predictive capacity of the trained perceptron-approximator MLP 7-7-1. Keywords: crown ethers, complexation constant, exploratory analysis, multiple linear regression, neural networks, modeling, prediction
DOI: 10.1134/S107036322010014X During the seminar on the occasion of 100th anniversary of J. Tukey, one of the founders of the practical data analysis [1], it has been stated that it was he who argued about reformation of the academic statistics, pointing at the existence of the unrecognized (at that time) branch of science aiming at the analysis of data [2]. The initial concepts and principles by J. Tukey has remained important and formed a basement of modern data science [3]. This statement is confirmed by numerous researches conducted over the recent decades in different fields of science and technology. It is recommended to pay extra attention to the data preparation and presentation prior to the analysis (Chambers, [4]) and to favor the predictive potential of mathematical models over mind deduction (Breiman, [5]). A comprehensive description of statistical and analysis methods applicable in science, industry, and business from the point of view of practical use is given in [6]. Predictive modeling opens incredible prospects for the theory development and acquiring new knowledge in the fields of science and technology dealing with big data analysis, especially in health science, ecology, chemistry, biology, and Earth science [7].
The approaches towards unsupervised and supervised pattern recognition, including methods of principal component analysis (PCA), nearest neighbor alog orithm (NN), discriminant analysis of partial least squares (PLS-DA), and artificial neural networks (ANN) [8]. Exploratory analysis is an important step following the data c
Data Loading...
