A machine learning system for detecting depression from audio and text inputs using advanced feature extraction and NLP techniques. Combines acoustic analysis (MFCC, spectral features) with linguistic pattern recognition for accurate predictions. Includes FastAPI REST API with real-time inference and interactive documentation.

This repository contains the scripts, results, and plots to reproduce the paper "Analyzing Temporal Factors for Anxiety and Depression Symptoms with the Rashomon Perspective".