Structural Analysis of Transition Metal β-X Substituent Interactions. Toward the Use of Soft Computing Methods for Catalyst Modeling

Abstract

Fuzzy logic and neural network techniques are used to classify intramolecular interactions between transition metals (M) and beta-X substituents in the following structural motif (LnMC(alpha)(A1)(A2)-C(beta)(B1)(B2)X). These interactions are relevant to the direct polymerization of functionalized olefins by Ziegler-Natta (ZN) catalysis. The efficiency and effectiveness of different soft computing techniques are compared. These methods give not only encouraging results with respect to general data mining issues but also insight into the factors that effect interactions between transition metals and beta-X substituents.

Modeling of chromatographic lipophilicity of food synthetic dyes estimated on different columns

“…Multivariate data analysis was performed using PCA incorporated in Statistica software, version 7.0 (http://www.statistica.com). Many problems from chemistry and other technical fields are strongly related to PCA 26–29. PCA is a typical projection method that allows one to find out the structure of data by depicting the objects (compounds) in a reduced coordinate system with maximum possible information.…”

Discrimination of haloarchaeal genera using Raman spectroscopy and robust methods for multivariate data analysis

“…Still, the major issue of the PCA algorithm remains embodied in the isolated points – outliers, extremes, etc. This can be handled by taking into account the outliers with respect to the first PC only . In this case, fuzzy membership degrees are defined according to the distance to the first PC.…”

“…The robust FPCA algorithm improves the PCA approach by fuzzification of the matrix data thus diminishing the influence of the outliers and poor linear correlation between variables. This results in a higher accounting for total variance, low‐dimensional matrix and more precise outline of principal components …”

In Silico Design in Homogeneous Catalysis Using Descriptor Modelling

“…This approach enables a simple and fast selection of the most promising catalysts candidates. Another application of neural networks and classification methods for data analysis in homogeneous catalysis is given by Cundari et al [86][87][88] The authors employed several data mining methods to disclose relationships between various metric parameters in transition metal imido complexes, a class of catalysts implicated in nitrogen fixation and C-H activation processes. The structures analysed were retrieved from the CSD and carried the motif: L n M=NZ, where M is the transition metal, L a ligand, N the nitrogen bound to the metal and Z a generic substituent.…”