Interface Access Behavior Modeling for Abnormal Request Interception
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing interface access control methods suffer from low interception accuracy, high manual costs, and long iteration cycles due to unauthorized access parties adapting to machine learning models, leading to information leakage and security vulnerabilities.
Innovation Solution
An identification model is used to analyze access action behavior sequences and time interval sequences to determine a classification probability for abnormal behaviors, enabling precise interception of interface access requests and reducing manual feature mining through end-to-end active learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional machine learning models are used for interface access control, then basic access monitoring can be achieved, but unauthorized access parties adapt to the models leading to low interception accuracy and information leakage
Solution Approach 1:
The patent implements dynamic model updating by continuously collecting new access behavior data and retraining the identification model. The system transitions from static to dynamic operation, where the model adapts to new attack patterns in real-time, preventing unauthorized parties from adapting to fixed model patterns and maintaining high interception accuracy.
Solution Approach 2:
The system establishes a feedback loop where interception results and access behaviors are continuously monitored and fed back into the model training process. This feedback mechanism allows the model to learn from both successful and failed interception cases, continuously improving its ability to detect unauthorized access while maintaining security reliability.
2Adaptability or versatility
If manual feature mining is used to build identification models, then model customization can be achieved, but the process incurs high manual costs and long iteration cycles
Solution Approach 1:
The system implements self-service through automated feature extraction and model training. Instead of requiring manual feature engineering, the system automatically extracts features from raw access behavior data and trains the identification model using machine learning algorithms. This automation eliminates manual intervention, reducing costs and accelerating iteration cycles while maintaining model adaptability to different access scenarios.
Solution Approach 2:
The patent replaces manual mechanical feature mining with automated computational feature extraction. Machine learning algorithms automatically identify and extract relevant features from access behavior data, substituting the manual mechanical process with an automated computational system that is both faster and more scalable.
3Reliability
If comprehensive access behavior analysis is performed to improve security, then unauthorized access detection can be enhanced, but system complexity and processing time increase
Solution Approach 1:
The system extracts and focuses on the most critical features from comprehensive access behavior data, such as access frequency, time patterns, and interface types. By selecting only the most relevant features for analysis, the system maintains high security detection capability while reducing computational complexity and processing time associated with analyzing all possible behavior parameters.
Data Source
AI summary
The present disclosure provides an interface access processing method and apparatus, a computer device and a storage medium, and the method includes: acquiring target access behavior sequences of an access party within a first specified time range, the target access behavior sequences including an access action behavior sequence and an access time interval sequence; calling an identification model for prediction processing to obtain an identification result based on the access action sequence and the access time interval sequence, the identification result being used for representing a classification probability that the access party has abnormal behaviors; and controlling an interface access request within a second specified time range of the access party according to the identification result, a start time of the second specified time range being later than an end time of the first specified time range.