FinInvest-GTCN: Explainable Graph-Temporal-Causal Modeling for Risk-Aware Investment Decision Optimization

Researchers introduce FinInvest-GTCN, a Graph-Temporal-Causal Network designed to optimize venture capital investment decisions by addressing challenges like heterogeneous data and non-stationary time series. The model redefines the task from content recommendation to quantitative risk-return assessment, utilizing a relational graph encoder, multi-scale temporal fusion, and a causal decision head to generate interpretable predictions.

• Combines a relational graph encoder for ecosystem topology, a multi-scale temporal fusion module for long-term dependencies, and a causal decision head for risk-adjusted predictions.
• Implements Meta-Causal Adaptation (MCA) to facilitate robust fine-tuning in data-scarce sectors by aligning updates with causally-plausible structures.
• Achieves state-of-the-art results on proprietary VC datasets, lowering Risk-Adjusted Mean Squared Error (RA-MSE) from 3.05 to 2.51.
• Boosts the cumulative return of a simulated portfolio by 18.7% compared to baselines.

This work pioneers a data-driven, explainable framework for investment decision support, offering enhanced stability and interpretability for high-stakes financial applications.