StackDPP: Stacking-Based Explainable Classifier for Depression Prediction and Finding the Risk Factors among...

Area of Science:

• Medical Informatics
• Computational Psychiatry
• Machine Learning in Healthcare

Background:

• Physician mental health is a critical concern globally.
• Identifying depression risk factors among physicians is challenging.
• Accurate prediction models are needed for timely intervention.

Purpose of the Study:

• To develop a machine learning-based predictive model for physician depression.
• To identify significant risk factors associated with physician depression.
• To evaluate the performance of various classification algorithms.

Main Methods:

• Collected and preprocessed physician health data.
• Utilized seven classification algorithms, including a novel stacking ensemble classifier (StackDPP).
• Tested models on 10 sub-datasets to optimize attribute selection.

Main Results:

• The proposed StackDPP model demonstrated superior performance across all datasets.
• Highest accuracy (0.962581) was achieved using all attributes.
• The top 20 attributes yielded accuracy (0.96129) comparable to using all attributes.
• Significant risk factors for physician depression were identified.

Conclusions:

• The StackDPP model effectively predicts depression levels in physicians.
• The model accurately identifies crucial risk factors for depression.
• Findings support enhanced treatment and therapy planning for physician mental health professionals.