Detecting Cognitive Distortions Through Machine Learning Text Analytics

Machine learning and text analytics have proven increasingly useful in a number of health-related applications, particularly in the context of analyzing online data for disease epidemics and warning signs of a variety of mental health issues. We follow in this tradition here, but focus our attention on cognitive distortion, a precursor and symptom of disruptive psychological disorders such as anxiety, anorexia and depression. We collected a number of personal blogs from the Tumblr API, and labeled them based on whether they exhibited distorted thought patterns. We then used LIWC to extract textual features and applied machine learning to the resulting vectors. Our findings show that it is possible to detect cognitive distortions automatically from personal blogs with relatively good accuracy (73.0%) and false negative rate (30.4%).

I. Introduction

With the explosion of social media data, health scientists have become more engaged in using computational modeling to better understand health and health behavior. Indeed, Pentland et al. have note that daily use of technology “leave digital breadcrumbs - tiny records of our daily experiences” that when mined and analyzed can provide insight into health behavior and health outcomes [26]. Similarly, a number of researchers have been advocating the value of online data in understanding public health problems and their determinants (e.g., see [17], [10], [15], [5], [21], [23], [24], [8]). It is clear that computer science expertise is critical to the collection and analysis of the associated data [29].