Researcher Muhannad Ahmed Amarneh, a student in the Master’s program in Data Science and Business Analytics, has defended his thesis titled “Predicting the Incidence of a Psychological Disorder (Anxiety, Depression and Stress) Using Machine Learning Algorithms.”
The present study aims to use machine learning algorithms to predict diagnoses of stress, anxiety and depression as the most common psychiatric disorders, using the dataset collected as part of this work. The dataset consisted of approximately 700 records using an online survey, which was based on the Depression, Anxiety and Stress International Scale (DASS21). The data was collected from Palestinian community participants and university students. To ensure the effectiveness of applying artificial intelligence algorithms, the data was processed before analyzing it. Thus, data was refined, duplicate data was searched for, some unnecessary elements were deleted, such as the hand used in writing and religion. Moreover, the texts were converted into numbers to ease using them with the algorithms. Because of the imbalance in the data, as some classifications contain a much larger number than others, rebalancing has been done to obtain more accurate results.
The researcher used five different algorithms to analyze the data and achieve early detection of mental health disorders. These algorithms are Random Forest, SVM, KNN, the XGBoost model and the Multi-Layer Perceptron (3layers) model. The results for depression SVM were the best in terms of model accuracy with 100%. Then, MLP with 98% followed with XGBoost with 95%, and finally KNN got a result of 79%. As for the anxiety results, SVM achieved the highest score of 100%, followed by MLP with 97%, both Random Forest and and XGBoost got a similar result of 96% and KNN got the lowest result of 78%.
The thesis was supervised by Professor Mohammed Awad. The committee of examiners included Dr. Ahmad Ewais and Dr. Youssef Daraghmeh.