This paper addresses the complexity and high-risk nature of fraud detection in financial transaction data by proposing a discrimination method based on an improved Mamba architecture. The study first performs data cleaning, normalization, and embedding to reduce scale differences across multimodal features, and employs temporal encoding and positional embedding to strengthen sequence representation. A dynamic state updating and selective state space mechanism is then introduced to capture long-term dependencies and global contextual features, thereby enhancing the model's discrimination ability in complex transaction environments. Furthermore, context interaction and weighted aggregation modules are designed to integrate hidden states across time steps into a global representation, providing more robust inputs for final classification. Experiments, including baseline comparisons, sequence length sensitivity, and noise robustness tests, demonstrate that the proposed method outperforms multiple mainstream architectures in Accuracy, F1-score, AUC, and Precision, while showing greater stability in handling data imbalance and adversarial perturbations. Overall, the study confirms the effectiveness and applicability of the improved Mamba architecture for financial fraud detection and provides technical support for building secure and trustworthy risk control systems.

This paper proposes a knowledge-graph-guided generative detection model to address the challenges of complex feature associations, hidden behavioral patterns, and sparse semantic information in financial fraud detection. The study maps multi-source heterogeneous data, such as accounts, transactions, devices, and geographic information, into a knowledge graph structure to model semantic dependencies and behavioral relationships among financial entities at both node and relation levels. Based on this, the model introduces a relation-aware encoder to obtain high-dimensional structural embeddings and employs a generative reasoning mechanism to learn the distribution of latent fraud patterns, enabling semantic identification of complex risk behaviors without relying on fixed rules. During generation, the method integrates knowledge constraints and structural priors, allowing the model to reconstruct potential fraud chains from a global semantic perspective and significantly enhance interpretability and generalization. Validation on real transaction datasets shows that the model outperforms comparison methods in AUC, ACC, Precision, and Recall, demonstrating the effectiveness of the knowledge-enhanced generative framework in capturing cross-entity dependencies and dynamic risk relationships. Overall, the proposed knowledge-graph-guided generative detection model achieves a unified framework of structural modeling, semantic reasoning, and anomaly generation, providing a new systematic solution for the intelligent detection of complex financial fraud behaviors. CCS CONCEPTS: Computing methodologies~Machine learning~Machine learning approaches.