Junghye Lee

Acquisition of a series of temperature-varied sample spectra to induce characteristic structural changes of components and selection of target-descriptive variables among them for multivariate analysis to improve accuracy

ABSTRACT As a means of improving the accuracy of Raman spectroscopic quantitative analysis, a strategy combining the generation of a series of temperature-varied spectra to make diverse and characteristic structural information of sample components widely available for calibration, and subsequent selection of more property-descriptive variables among these spectra, has been demonstrated. For the evaluation, Raman spectra of synthetic hydrocarbon mixtures, lube base oils (LBOs) and polyethylene (PE) pellets were acquired at regular intervals while the sample temperature gradually increased from cryogenic to near room temperature. To select target-descriptive variables from all of the snapshot (temperature-varied) spectra, a Markov blanket (MB) feature (variable) selection able to produce a minimal set of features without changing the original target distribution was adopted. The selection utilizes a conditional independence test to quickly obtain an optimal feature subset by simultaneously considering relevance to a target variable and redundancy between selected features without using any heuristic searching. When MB-selected variables were used for partial least squares (PLS) to determine the concentrations of components in the hydrocarbon mixtures, kinematic viscosity of 40°C (KV@40) of LBOs, and density of PE pellets, the accuracy was improved compared to the use of either all snapshot spectra without variable selection or an optimal single snapshot spectrum. The incorporation of component-specific and property-descriptive variables without redundant information for PLS was the origin of the improvement in accuracy. The proposed method could potentially be extended to the analysis of other complex samples including petroleum-driven samples, edible oils, and other polymers.

Interleaved incremental association Markov blanket as a potential feature selection method for improving accuracy in near-infrared spectroscopic analysis

The interleaved Incremental Association Markov Blanket (inter-IAMB) is described herein as a feature selection method for the NIR spectroscopic analysis of several samples (diesel, gasoline, and etchant solutions). Although the Markov blanket (MB) has been proven to be the minimal optimal set of features (variables) that does not change the original target distribution, variables selected by the existing IAMB algorithm could be redundant and/or misleading as the IAMB requires an unnecessarily large amount of learning data to identify the MB. Use of the inter-IAMB interleaving the grow phase with the shrink phase to maintain the size of the MB as small as possible by immediately eliminating invalid candidates could overcome this drawback. In this report, a likelihood-ratio (LR)-based conditional independence test, able to handle spectroscopic data normally comprising a large number of continuous variables in a small number of samples, was uniquely embedded in the inter-IAMB and its utility was evaluated. The variables selected by the inter-IAMB in complexly overlapped and feature-indistinct NIR spectra were used to determine the corresponding sample properties. For comparison, the properties were also determined using the IAMB-selected variables as well as the whole variables. The inter-IAMB was more effective in the selection of variables than the IAMB and thus able to improve the accuracy in the determination of the sample properties, even though a smaller number of variables was used. The proposed LR-embedded inter-IAMB could be a potential feature selection method for vibrational spectroscopic analysis, especially when the obtained spectral features are specificity-deficient and extensively overlapped.