Predicting Negative Emotions Based on Mobile Phone Usage Patterns: An Exploratory Study

Abstract

BackgroundPrompt recognition and intervention of negative emotions is crucial for patients with depression. Mobile phones and mobile apps are suitable technologies that can be used to recognize negative emotions and intervene if necessary.ObjectiveMobile phone usage patterns can be associated with concurrent emotional states. The objective of this study is to adapt machine-learning methods to analyze such patterns for the prediction of negative emotion.MethodsWe developed an Android-based app to capture emotional states and mobile phone usage patterns, which included call logs (and use of apps). Visual analog scales (VASs) were used to report negative emotions in dimensions of depression, anxiety, and stress. In the system-training phase, participants were requested to tag their emotions for 14 consecutive days. Five feature-selection methods were used to determine individual usage patterns and four machine-learning methods were tested. Finally, rank product scoring was used to select the best combination to construct the prediction model. In the system evaluation phase, participants were then requested to verify the predicted negative emotions for at least 5 days.ResultsOut of 40 enrolled healthy participants, we analyzed data from 28 participants, including 30% (9/28) women with a mean (SD) age of 29.2 (5.1) years with sufficient emotion tags. The combination of time slots of 2 hours, greedy forward selection, and Naïve Bayes method was chosen for the prediction model. We further validated the personalized models in 18 participants who performed at least 5 days of model evaluation. Overall, the predictive accuracy for negative emotions was 86.17%.ConclusionWe developed a system capable of predicting negative emotions based on mobile phone usage patterns. This system has potential for ecological momentary intervention (EMI) for depressive disorders by automatically recognizing negative emotions and providing people with preventive treatments before it escalates to clinical depression.

Factors associated with mental health among high school students in Iran: Does mobile phone overuse associate with poor mental health?

“…Our results also showed that gender, education grade, the number of daily received/sent text messages, and being awakened at night to use MP were significant predictors of poor mental health. Similar with those found in previous studies (Augner & Hacker, ; Hung, Yang, Chang, Chiang, & Chen, ; Kowalski & Limber, ; Santini et al., ), female students with MP overuse had 1.5 times poorer mental health compared to the male students. In some other studies, despite more excessive level of MP use among male students, in proportion to the female counterparts, there was no significant level of poor mental health among the male students (Bianchi & Phillips, ; Mazaheri et al., ; Zulkefly & Baharudin, ).…”

Automatic detection of emotions in depressed patients and healthy subjects: a mobile application study (Preprint)

“…Note that these 2n values were individually chosen based on the distribution of each participant's data. Accordingly, for the variables "Sleep quality", "Total time", 11 and "Total frequency", they were discretized into "H" and "L". For each application category, we combined discretized usage time and frequency to four attributes: "HH", "LH", "LL", and "HL".…”

“…Comparatively more studies were conducted in healthy population and found that location variability, environmental features (e.g. ambient light and sound), physical activity, sleep duration and phone use patterns are correlated with depressive symptoms [8][9][10][11][12]. Emotion-aware mobile applications have been developed using complex machine learning methods [11,[13][14][15].…”

“…ambient light and sound), physical activity, sleep duration and phone use patterns are correlated with depressive symptoms [8][9][10][11][12]. Emotion-aware mobile applications have been developed using complex machine learning methods [11,[13][14][15]. Currently it is unclear which type of information is universally crucial for detecting emotion; alternatively, there is evidence that the detection algorithm must be personalized [9,11].…”

“…Emotion-aware mobile applications have been developed using complex machine learning methods [11,[13][14][15]. Currently it is unclear which type of information is universally crucial for detecting emotion; alternatively, there is evidence that the detection algorithm must be personalized [9,11]. Moreover, certain machine-learning methods function like a 'black box', such that contributors to an emotional detection can't be identified.…”