Mental Health Classifier

Project Type: AI/ML · Binary Classification

Overview

This project explores the impact of remote work on mental health by training a neural network to predict depression based on survey responses. The dataset includes demographic and work-related features, all of which are encoded and used to train a PyTorch-based model.

Motivation

As remote work continues to grow, there’s a growing concern about its long-term impact on employee well-being. This project aims to understand which factors most significantly correlate with mental health conditions like depression and whether we can reliably predict risk using basic survey data. The goal is to use machine learning not just for prediction, but also to inform better workplace strategies and support systems.

Approach

• Exploratory Data Analysis (EDA) to understand the structure and trends in the dataset
• Encoded categorical variables using pandas.factorize()
• Split the dataset into train/test using train_test_split() from Scikit-learn
• Constructed a simple fully connected neural network using PyTorch
• Trained the model with binary cross-entropy loss and Adam optimizer
• Evaluated accuracy and logged performance results for insights

Technologies Used

Python, Pandas, PyTorch, Scikit-learn

🔗 View GitHub Repository

← Back to Portfolio