Fantastic Features and Where to Find Them: Detecting Cognitive Impairment with a Subsequence Classification Guided Approach

TL;DR

This paper introduces a novel feature engineering approach that combines sequential models and domain knowledge to improve cognitive impairment detection from speech, achieving a 2.3% accuracy boost over baseline methods.

Contribution

The paper presents a new method for selecting engineered features using sequential models guided by domain knowledge, enhancing interpretability and performance in health-related NLP tasks.

Findings

Improved classification accuracy by 2.3% using the proposed feature selection method.

Demonstrated effectiveness on a standard cognitive impairment speech dataset.

Showcased the method's potential for interpretability in healthcare applications.

Abstract

Despite the widely reported success of embedding-based machine learning methods on natural language processing tasks, the use of more easily interpreted engineered features remains common in fields such as cognitive impairment (CI) detection. Manually engineering features from noisy text is time and resource consuming, and can potentially result in features that do not enhance model performance. To combat this, we describe a new approach to feature engineering that leverages sequential machine learning models and domain knowledge to predict which features help enhance performance. We provide a concrete example of this method on a standard data set of CI speech and demonstrate that CI classification accuracy improves by 2.3% over a strong baseline when using features produced by this method. This demonstration provides an ex-ample of how this method can be used to assist classification in fields where interpretability is important, such as health care.