Synthetic Mental Health Survey Dataset

The Synthetic Mental Health Prediction Dataset was developed  as part of a university thesis project on machine learning\u2013based mental health analysis. The dataset is designed to simulate survey-based responses for predicting the severity of six mental health disorders, including Depression, Anxiety, ADHD, PTSD, Bipolar Disorder, and Substance Use, as well as assessing overall mental health status.The dataset contains demographic information such as Name, Gender, Age, Occupation, and Living Situation. It also includes responses to 20 weighted survey questions, organized into six clinically significant domains: Emotional Well-being (30%), Cognitive Functioning (20%), Behavioral Indicators (15%), Physical Symptoms (15%), History and Risk Factors (10%), and Substance Use (10%). Each question has an assigned clinical weight, which allows for the calculation of scores at both the category level and the total level.The data file consists of 27 input features, including the 20 individual question scores, six category sums, and the total score. It also contains 10 output targets representing disorder severity as continuous values and overall mental health status as a binary classification (Good, Average, Bad, Very Bad). All data are synthetically generated, anonymized, and ethically compliant, intended exclusively for academic, analytical, and machine learning research purposes.This dataset is suitable for studies in predictive modeling, multi-output regression, and classification, and it can be used to explore early mental health assessment using machine learning algorithms such as Random Forest, XGBoost, and LightGBM.