Hybrid DeepSentX Framework for AI-Driven Requirements Insight and Risk Prediction in Multilingual Sports Using...

Area of Science:

• Artificial Intelligence
• Natural Language Processing
• Software Engineering

Background:

• Engineering teams require real-time signals for evolving requirements and release risks.
• Multilingual fan discourse in live sports is characterized by noise, code-switching, sarcasm, and event-driven drift.

Purpose of the Study:

• To present Hybrid DeepSentX, an AI-driven framework for converting crowd commentary into actionable requirements insight and sprint-level risk scores.
• To enhance the robustness of AI models against sarcasm and event-driven drift in multilingual online discussions.

Main Methods:

• Coupling multilingual transformer encoders with an inductive GraphSAGE conversation graph.
• Integrating a reinforcement learner with rewards shaped for sarcasm and rapidly shifting events.
• Utilizing a million-plus post corpus from X, Reddit, and sports forums for evaluation.

Main Results:

• Hybrid DeepSentX achieved higher macro-averaged accuracy and F1 scores on code-switched and sarcastic subsets compared to baselines.
• The framework demonstrated reduced missed risk flags and generated developer-facing artifacts for backlog grooming and defect triage.
• Achieved superior performance in accuracy, calibration, latency, memory, and parameter count.

Conclusions:

• Hybrid DeepSentX offers a unified multilingual design integrating transformer, graph, and reinforcement components for robust analysis of online discourse.
• The framework provides valuable deployment artifacts, transforming model outputs into requirement clusters and sprint risk scores for continuous planning.
• The study introduces a novel annotated multi-platform corpus with explicit code-switching and sarcasm labels.