Dynamic Forecasting and Temporal Feature Evolution of Stock Repurchases in Listed Companies Using Attention-Based Deep Temporal Networks

TL;DR

This paper introduces a deep learning-based dynamic forecasting system for stock repurchases, capturing complex temporal patterns and providing explainable insights into corporate financial decision triggers.

Contribution

It develops a hybrid deep temporal network model integrating TCN and attention mechanisms, advancing stock repurchase prediction with explainability and empirical validation.

Findings

Model significantly outperforms static baselines like Logistic Regression and XGBoost.

Prolonged undervaluation is identified as a long-term motive for repurchases.

Sharp increases in cash flow act as short-term triggers for repurchase decisions.

Abstract

Accurately predicting stock repurchases is crucial for quantitative investment and risk management, yet traditional static models fail to capture the complex temporal dependencies of corporate financial conditions. This paper proposes a dynamic early warning system integrating economic theory with deep temporal networks. Using Chinese A-share panel data (2014-2024), we employ a hybrid Temporal Convolutional Network (TCN) and Attention-based LSTM to capture long- and short-term financial evolutionary patterns. Rolling-window cross-validation demonstrates our model significantly outperforms static baselines like Logistic Regression and XGBoost. Furthermore, utilizing Explainable AI (XAI), we reveal the temporal dynamics of repurchase decisions: prolonged "undervaluation" serves as the long-term underlying motive, while a sharp increase in "cash flow" acts as the decisive short-term trigger. This study provides a robust deep learning paradigm for financial forecasting and offers dynamic empirical support for classic corporate finance hypotheses.