Use of a neural network to determine the boiling point of alkanes

Cherqaoui, Driss; Villemin, Didier

doi:10.1039/ft9949000097

Cited by 85 publications

(94 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This kernel is more likely to capture the global structure of the molecules and their chemical groups. On the prediction of the boiling point of alkanes, the 1D, 2D, and 3D histogram kernels achieve results comparable to those in Cherqaoui and Villemin 38 and Micheli et al 39 This is consistent with their observation that most of the relevant information is captured by the knowledge of how many carbons are attached to each carbon. On the Bergström data set, using leave-one-out cross-validation, we obtain an improved squared correlation of 0.69, versus the 0.63 value reported by Bergström et al Detailed results on the small data sets are reported in Table 1.…”

Section: Small Data Setssupporting

confidence: 83%

One- to Four-Dimensional Kernels for Virtual Screening and the Prediction of Physical, Chemical, and Biological Properties

Azencott

Ksikes

Swamidass

et al. 2007

J. Chem. Inf. Model.

View full text Add to dashboard Cite

Many chemoinformatics applications, including high-throughput virtual screening, benefit from being able to rapidly predict the physical, chemical, and biological properties of small molecules to screen large repositories and identify suitable candidates. When training sets are available, machine learning methods provide an effective alternative to ab initio methods for these predictions. Here, we leverage rich molecular representations including 1D SMILES strings, 2D graphs of bonds, and 3D coordinates to derive efficient machine learning kernels to address regression problems. We further expand the library of available spectral kernels for small molecules developed for classification problems to include 2.5D surface and 3D kernels using Delaunay tetrahedrization and other techniques from computational geometry, 3D pharmacophore kernels, and 3.5D or 4D kernels capable of taking into account multiple molecular configurations, such as conformers. The kernels are comprehensively tested using cross-validation and redundancy-reduction methods on regression problems using several available data sets to predict boiling points, melting points, aqueous solubility, octanol/water partition coefficients, and biological activity with state-of-the art results. When sufficient training data are available, 2D spectral kernels in general tend to yield the best and most robust results, better than state-of-the art. On data sets containing thousands of molecules, the kernels achieve a squared correlation coefficient of 0.91 for aqueous solubility prediction and 0.94 for octanol/water partition coefficient prediction. Averaging over conformations improves the performance of kernels based on the three-dimensional structure of molecules, especially on challenging data sets. Kernel predictors for aqueous solubility (kSOL), LogP (kLOGP), and melting point (kMELT) are available over the Web through: http:// cdb.ics.uci.edu.

show abstract

Section: Small Data Setssupporting

confidence: 83%

One- to Four-Dimensional Kernels for Virtual Screening and the Prediction of Physical, Chemical, and Biological Properties

Azencott

Ksikes

Swamidass

et al. 2007

J. Chem. Inf. Model.

View full text Add to dashboard Cite

show abstract

“…A Recursive Neural Networks approach, based on Recursive Cascade Correlation (RCC) model [2], has been proved to be competitive with respect to ad-hoc techniques at the state-of-art in the field. In particular, the results have been compared with the approach reported in [3], which yields very good performances applying a multilayered feedforward neural network to a vectorial representation of alkanes able to retain the structural information which is known to be relevant to the prediction of the boiling point.…”

Section: Application To Qspr Analysismentioning

confidence: 99%

“…In Table 1 we give the performance of the Tree Fisher Score methods (TFK-CG refers to the version with Gaussian variables) compared with the Recursive Cascade Correlation (RCC) employed in [2] and the ad-hoc Multilayered Perceptron employed in [3].…”

Section: A Overview Of Experimentsmentioning

confidence: 99%

“…The data set that composes our benchmark, which is taken from [3], is based on all the 150 alkanes with up to 10 carbon atoms (C n H 2n+2 ). It is well known that for this class of compounds, the prediction of the boiling point can be performed by disregarding the information about the hydrogen atoms.…”

Section: Application To Qspr Analysismentioning

confidence: 99%

“…predicting the boiling point of a group of alkanes, to test the quality of the parameters extracted by the Tree Fisher Kernel. This is a regression problem on a structured domain involving chemical compounds represented as trees and it has been used in testing various QSPR approaches [2], [3].…”

Section: Application To Qspr Analysismentioning

confidence: 99%

See 2 more Smart Citations

Fisher kernel for tree structured data

Nicotra

Micheli²,

Starita³

2004 IEEE International Joint Conference on Neural Networks (IEEE Cat. No.04CH37541)

View full text Add to dashboard Cite

Abstract-We introduce a kernel for structured data, which is an extension of the Fisher Kernel used for sequences [11]. In our approach, we extract the Fisher score vectors from a Bayesian Network, specifically a Hidden Tree Markov Model [6], which can be constructed starting from the training data. Experiments on a QSPR (quantitative structure-property relationship) analysis, where instances are naturally represented as trees, allow a first test of the approach.

show abstract