Pre-Detecting Risky Code via Multi-Dimensional User and Keyword Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing code risk control systems are not timely in detecting and mitigating risks, leading to potential security breaches and significant losses.
Innovation Solution
A method and apparatus for pre-detecting risky codes by analyzing a user's historical evaluation records, login times, and login addresses, combined with keyword analysis, to determine a risk level before code execution, using preset rules to calculate trustworthiness, time consistency, and regional scores, and integrating these scores to assess the risk level of the code.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If code risk detection is performed after code execution, then the detection process is simpler, but security breaches and losses occur before detection can happen
Solution Approach 1:
The system performs preliminary risk detection by analyzing historical evaluation records, login times, login addresses, and code keywords before the code is executed. This advance analysis generates a risk level that determines whether the code should be allowed to execute, thereby preventing security breaches before they occur rather than detecting them after execution.
2Measurement precision
If multiple historical data dimensions are analyzed to improve risk identification accuracy, then detection precision improves, but system complexity increases
Solution Approach 1:
The risk detection system is segmented into four independent analysis modules: historical evaluation record analysis, login time analysis, login address analysis, and code keyword analysis. Each module processes a specific dimension of data and generates a separate score, which are then combined to determine the overall risk level. This segmentation improves detection precision through comprehensive analysis while managing system complexity through modular design.
Data Source
AI summary
A method for pre-detecting a risky code, which is performed by an apparatus for pre-detecting the risky code. The method includes: receiving a code to be detected input by a target user, and obtaining a target historical evaluation record of the target user, a target historical login time and a target historical login address for login to the apparatus for pre-detecting the risky code: analyzing the target historical evaluation record based on a first preset rule to obtain a target trustworthiness score, analyzing the target historical login time based on a second preset rule to obtain a time consistency score, analyzing the target historical login address based on a third preset rule to obtain a regional score, analyzing keyword information of the code to be detected based on a fourth preset rule to obtain a keyword score; and analyzing the target user trustworthiness score, the time consistency score, the regional score, and the keyword score to obtain a risk level of the code to be detected, and outputting the risk level.