An online business interaction anomaly analysis method based on an AI server and big data

By constructing a two-way interactive communication network and a multimodal interactive feature network, interactive features are acquired and deep interactive feature profiles are created, solving the problem of cross-dimensional anomaly detection in traditional methods and improving the analysis level of online business interactions and the processing power of AI servers.

CN121000773BActive Publication Date: 2026-06-26LOGOSDATA
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
LOGOSDATA
Filing Date
2025-09-12
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Traditional methods for analyzing anomalies in online business interactions cannot effectively detect anomalies that cross dimensions, and the insufficient computing power of AI servers leads to latency, making it difficult to accurately analyze anomalies in complex online business interaction scenarios.

Method used

Construct a two-way interactive communication network, build a multimodal interactive feature network through multimodal cross-interfaces, acquire interactive features and perform deep interactive feature profiling, and combine contextual feature profiling to detect abnormal online business interaction users.

Benefits of technology

It enables cross-dimensional analysis of online business interactions, improves the accuracy of anomaly detection and the processing power of AI servers, and ensures the stability and security of business interactions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121000773B_ABST
    Figure CN121000773B_ABST
Patent Text Reader

Abstract

The application discloses an online business interaction abnormality analysis method based on an AI server and big data, and relates to the technical field of AI servers and big data; the application constructs a bidirectional interaction communication network, sets a multi-modal cross interface, constructs a multi-modal interaction feature network of the bidirectional interaction communication network through the multi-modal cross interface, acquires interaction features, and acquires a multi-modal interaction data chain of online business interaction of the interaction features through the multi-modal interaction feature network; the depth interaction feature portrait of the interaction features is acquired through the multi-modal interaction data chain; the depth interaction story is acquired by performing context feature portrait fusion on the depth interaction feature portrait; and the abnormal online business interaction user is acquired by detecting the depth interaction story; the application realizes the analysis level of the cross-dimension analysis of online business interaction.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Abnormality judgment method and device, electronic equipment and storage medium

    CN116244425A

  • Abnormal behavior detection method based on multi-modal fusion and interpretability analysis

    CN120086753A