http://romisatriawahono.net/lecture/rm/survey/machine%20learning/Chandrashekar%20-%20Feature%20Selection%20Methods%20-%202014.pdf

A survey on feature selection methods

This paper emphasizes the importance of rigorous validation as a crucial, integral component of Quantitative Structure Property Relationship (QSPR) model development. We consider some examples of published QSPR models, which in spite of their high fitted accuracy for the training sets and apparent mechanistic appeal, fail rigorous validation tests, and, thus, may lack practical utility as reliable screening tools. We present a set of simple guidelines for developing validated and predictive QSPR models. To this end, we discuss several validation strategies including (1) randomization of the modelled property, also called Y-scrambling, (2) multiple leave-many-out cross-validations, and (3) external validation using rational division of a dataset into training and test sets. We also highlight the need to establish the domain of model applicability in the chemical space to flag molecules for which predictions may be unreliable, and discuss some algorithms that can be used for this purpose. We advocate the broad use of these guidelines in the development of predictive QSPR models.

The Importance of Being Earnest: Validation is the Absolute Essential for Successful Application and Interpretation of QSPR Models

The Mahalanobis distance

A review of near infrared spectroscopy and chemometrics in pharmaceutical technologies.

The successive projections algorithm for variable selection in spectroscopic multicomponent analysis

Near-infrared spectroscopy applications in pharmaceutical analysis

A comparison of multiple linear regression (MLR) with partial least-squares (PLS) regression is presented, for the multivariate modeling of hydroxyl number in a certain polymer of a heterogeneous near-IR spectroscopic data set. The MLR model was performed by selecting the variables with a genetic algorithm. A good model could be obtained with both methods. It was shown that the MLR and PLS solutions are very similar. The two models include the same number of variables, and the first variables in each model have similar, chemically understandable functions. It is concluded that both solutions are equivalent and that each has some advantages and disadvantages. This also means that even in very complex situations such as here, MLR can replace PLS in certain cases.

D.L. Massart

Papers

The Mahalanobis distance

Near-infrared spectroscopy applications in pharmaceutical analysis

Genetic Algorithms as a Tool for Wavelength Selection in Multivariate Calibration

The influence of data pre-processing in the pattern recognition of excipients near-infrared spectra.

Assessment of new constraints applied to the alternating least squares method