This document proposes using deep learning techniques like LSTM neural networks to predict corporate bankruptcy by integrating both financial ratio data and textual disclosures from annual reports. It notes that previous studies have largely relied on statistical models or used only financial data with machine learning. The researcher aims to determine if adding textual data to an LSTM model improves prediction performance over a CNN model using only financial ratios. The document outlines the research question, objectives, and provides an overview of previous bankruptcy prediction studies using statistical, machine learning and deep learning methods.