I know people who talk to their computer, and not in a SIRI-kind of way. I know other people who think their computer holds a grudge against them. But what if computers would know how you feel? Researchers now have designed a computer program that can accurately recognize users’ emotional states as much as 87% of the time, depending on the emotion. The study combined — for the first time — two established ways of detecting user emotions: keystroke dynamics and text-pattern analysis.

From the press release:

Writing in the journal Behaviour & Information Technology, A.F.M. Nazmul Haque Nahin and his colleagues describe how their study combined — for the first time — two established ways of detecting user emotions: keystroke dynamics and text-pattern analysis.

To provide data for the study, volunteers were asked to note their emotional state after typing passages of fixed text, as well as at regular intervals during their regular (‘free text’) computer use; this provided the researchers with data about keystroke attributes associated with seven emotional states (joy, fear, anger, sadness, disgust, shame and guilt). To help them analyse sample texts, the researchers made use of a standard database of words and sentences associated with the same seven emotional states.

After running a variety of tests, the researchers found that their new ‘combined’ results were better than their separate results; what’s more, the ‘combined’ approach improved performance for five of the seven categories of emotion. Joy (87%) and anger (81%) had the highest rates of accuracy.

This research is an important contribution to ‘affective computing’, a growing field dedicated to ‘detecting user emotion in a particular moment’. As the authors note, for all the advances in computing power, performance and size in recent years, a lot more can still be done in terms of their interactions with end users. “Emotionally aware systems can be a step ahead in this regard,” they write.

“Computer systems that can detect user emotion can do a lot better than the present systems in gaming, online teaching, text processing, video and image processing, user authentication and so many other areas where user emotional state is crucial.”

While much work remains to be done, this research is an important step in making ’emotionally intelligent’ systems that recognise users’ emotional states to adapt their music, graphics, content or approach to learning a reality.

Abstract of the research:

Emotion is a cognitive process and is one of the important characteristics of human beings that makes them different from machines. Traditionally, interactions between humans and machines like computers do not exhibit any emotional exchanges. If we could build any system that is intelligent enough to interact with humans that involves emotions, that is, it can detect user emotions and change its behaviour accordingly, then using machines could be more effective and friendly. Many approaches have been taken to detect user emotions. Affective computing is the field that detects user emotion in a particular moment. Our approach in this paper is to detect user emotions by analysing the keyboard typing patterns of the user and the type of texts (words, sentences) typed by them. This combined analysis gives us a promising result showing a substantial number of emotional states detected from user input. Several machine learning algorithms were used to analyse keystroke timing attributes and text pattern. We have chosen keystroke because it is the cheapest and most available medium to interact with computers. We have considered seven emotional classes for classifying the emotional states. For text pattern analysis, we have used vector space model with Jaccard similarity method to classify free-text input. Our combined approach showed above 80% accuracies in identifying emotions.