Cross-validation of bimodal health-related stress assessment

Abstract

This chapter is the second of the set of two chapters that aim towards bringing affective computing to practice. The previous chapter described two studies with the aim to employ ASP to initialize Computer-Aided Diagnosis (CAD) for mental health care. This chapter continues this endeavor with the data of these two studies. Two instruments were chosen to assess the stress level of the patients at various point in time during therapy: I) speech, used as an objective and ubiquitous stress indicator, and II) the Subjective Unit of Distress (SUD), a clinically validated Likert scale. In total, 13 statistical parameters were derived from each of 5 speech features: amplitude, zero crossings, power, high-frequency power, and pitch. To model the emotional state of the patients, 28 parameters were selected from this set by means of a linear regression model presented in the previous chapter. Subsequently, this representation was compressed into 11 principal components. The SUD and speech model were cross-validated, using 3 machine learning techniques (i.e., K Nearest Neighbors, Support Vector Machines, and Multi-Layer Perceptron neural network). Between 90% (2 SUD levels) and 39% (10 SUD levels) correct classification was achieved. The two sessions could be discriminated in 89% (for ST) and 77% (for RL) of the cases. This report fills a gap between laboratory and clinical studies, as presented in the previous chapter, and its results emphasize the usefulness of Computer Aided Diagnostics (CAD) for mental health care.

This chapter is based on:

Broek, E.L. van den, Sluis, F. van der, and Dijkstra, T. Cross-validation of bimodal healthrelated stress assessment. Personal and Ubiquitous Computing. [in press]