Principal components variable importance reconstruction (PC-VIR): Exploring predictive importance in multicollinear acoustic speech data

TL;DR

This paper introduces PC-VIR, a method for assessing variable importance in multicollinear data, validated on acoustic speech features, showing comparable or improved predictive performance over existing methods.

Contribution

The paper presents PC-VIR, a novel approach for evaluating variable importance in multicollinear datasets, with a focus on acoustic speech data, incorporating Bonferroni correction for significance testing.

Findings

PC-VIR provides conservative importance estimates similar to logistic regression.

The method achieves comparable data fit to partial least squares regression.

Predictions using PC-VIR are as accurate as those from partial least squares methods.

Abstract

This paper presents a method of exploring the relative predictive importance of individual variables in multicollinear data sets at three levels of significance: strong importance, moderate importance, and no importance. Implementation of Bonferroni adjustment to control for Type I error in the method is described, and results with and without the correction are compared. An example of the method in binary logistic modeling is demonstrated by using a set of 20 acoustic features to discriminate vocalic nasality in the speech of six speakers of the Mixean variety of Low Navarrese Basque. Validation of the method is presented by comparing the direction of significant effects to those observed in separate logistic mixed effects models, as well as goodness of fit and prediction accuracy compared to partial least squares logistic regression. The results show that the proposed method yields: (1) similar, but more conservative estimates in comparison to separate logistic regression models, (2) models that fit data as well as partial least squares methods, and (3) predictions for new data that are as accurate as partial least squares methods.