An Interactive Mobile Computing and Analytical Framework for Complex Financial Data with Applications to Audit Tasks

Abstract

The growing need for real-time interactive analysis of financial data in mobile on-site auditing is hindered by mobile computational limits and poor network conditions, leading to high multidimensional query latency, low fraud detection accuracy, and weak data synchronization reliability. Existing end-edge-cloud frameworks are not tailored to financial auditing, failing to simultaneously meet real-time interactivity, low terminal power consumption, and high-precision risk detection. In this study, an end-edge-cloud three-layer collaborative interactive mobile computing and analytical framework was constructed for financial auditing, in which dynamic and coordinated scheduling of multi-tier resources was achieved through a task offloading agent. To address domain-specific challenges, a set of tailored optimization mechanisms was proposed: multidimensional financial data query efficiency on mobile devices was enhanced through user behavior prediction and incremental visualization strategies; a two-stage inference architecture combining coarse-grained screening on the end side and fine-grained judgment on the edge side was designed, where global cost optimization and proportional-integral-derivative-based adaptive threshold control were integrated to achieve a dynamic balance between detection accuracy and device energy consumption; a multiconstrained ternary task offloading model and a dependency-aware scheduling strategy were developed to accommodate the tidal workload patterns of audit tasks; a network-graded differential synchronization mechanism was established to guarantee the consistency of financial data synchronization under weak network conditions; and for core audit scenarios, a dedicated algorithmic system incorporating heterogeneous anomaly scoring, graph convolutional fund flow mining, and dual-encoder evidence matching was constructed. This study advances the theory of mobile edge collaborative computing in the financial auditing domain and provides an efficient and feasible technical solution for the large-scale deployment of intelligent mobile auditing.