1. Introduction

Mental health is a crucial aspect of overall well-being, yet diagnosing and managing mental health conditions remain challenging. This project explores the use of machine learning models for classifying mental health severity using socio-demographic and behavioral data.

Objective: The aim is to evaluate and compare the performance of different machine learning models in predicting the severity of mental health conditions.

2. Dataset

Source: The dataset used in this study consists of 1,000 data points covering socio-demographic and behavioral attributes.

Target: The target variable represents the severity of mental health conditions categorized into four levels: None, Mild, Moderate, and Severe.

2.1 Dataset Features

3. Data Preprocessing

3.1 Data Cleaning

3.2 Dimensionality Reduction

3.3 Addressing Class Imbalance

4. Model Development

4.1 Algorithms Used

4.2 Model Training

  1. Data Split: The dataset was split into 80% training and 20% testing.
  2. Hyperparameter Tuning: Grid search was applied to optimize hyperparameters for SVM and LightGBM.
  3. Cross-Validation: A 5-fold cross-validation approach was implemented to enhance model generalization.

4.3 Evaluation Metrics

5. Results and Comparison

Model Accuracy Precision Recall F1-Score
Logistic Regression 50.3% 0.49 0.50 0.46
LightGBM 51.0% 0.52 0.52 0.48
SVM with PCA+SMOTE 55.5% 0.56 0.56 0.55

Key Observations:

6. Conclusion

This study highlights the effectiveness of machine learning in classifying mental health conditions. The SVM model with PCA and SMOTE outperformed other models, demonstrating the value of dimensionality reduction and class balancing techniques.

Future Work

7. How to Run the Project

Prerequisites

Steps

  1. Clone the repository:

    git clone https://github.com/username/mental-health-classification.git
    
  2. Run the Jupyter notebook:

    jupyter notebook mental_health_classification.ipynb