stock crash prediction dataset in korean market

This dataset contains stock-level features and target variables used for predicting individual stock crash risk in the Korean equity market from 2015 to 2024.It includes both financial indicators and volatility-related metrics, such as VKOSPI, and is designed for use in machine learning research on downside risk forecasting.Each record represents a specific stock and trading date, with the binary label WindowDrop indicating whether a price drop exceeding 40% occurred within the following 60 trading days.This dataset supports the findings of the paper \u201cStock Crash Risk Prediction with Implied Volatility Index: A Comparison of Tree-Based and Transformer-Based Models\u201d