Malicious Account Identification via Reading Speed Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying malicious accounts in net literature platforms, which engage in fake book viewing, are ineffective due to the manual setting of thresholds without a theoretical basis, allowing cheaters to easily evade detection and distort the normal operation of these platforms.
Innovation Solution
A method and apparatus that utilize a server to determine a reading speed threshold based on distribution data from a known set of normal accounts, using effective reading speed as a training sample and applying a Gaussian distribution model to identify accounts with reading speeds exceeding the threshold as malicious, with additional assistance from external data for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a fixed threshold of maximum read words is manually set, then it is easier to operate and implement, but the identification accuracy of malicious accounts deteriorates because cheaters can easily evade detection by limiting their fake reading below the threshold
Solution Approach 1:
The patent changes the parameter from a fixed manually-set threshold to a dynamically determined threshold based on statistical distribution of reading speeds. The system calculates the upper confidence bound of the reading speed distribution for normal users, and uses this as the threshold. This transforms the threshold from a static parameter to one that adapts based on observed data, thereby improving detection accuracy while maintaining operational simplicity.
Solution Approach 2:
The patent replaces the manual mechanical process of setting thresholds with an automated statistical computation system. Instead of operators manually adjusting thresholds based on experience, the system automatically computes the threshold using Gaussian distribution fitting and confidence interval calculation on the collected reading speed data, eliminating the need for manual intervention and improving both accuracy and consistency.
2Reliability
If a large fixed threshold is set to reduce incorrect determinations, then fewer normal accounts are mistakenly identified as malicious, but the ability to detect actual malicious accounts deteriorates because malicious accounts can operate below this high threshold
Solution Approach 1:
The patent introduces dynamics into the threshold determination process. Instead of using a static large threshold, the system continuously monitors reading speed data and dynamically adjusts the threshold based on the observed distribution. The threshold is set at the upper confidence bound, which automatically adapts to changes in normal user behavior patterns, thereby maintaining high reliability while improving detection capability.
Solution Approach 2:
The system implements feedback by continuously collecting reading speed data, recalculating the distribution parameters, and adjusting the threshold accordingly. This closed-loop approach ensures that the threshold remains optimized for detecting malicious accounts while minimizing false positives, as the system learns from actual user behavior patterns over time.
3Device complexity
If manual threshold setting is used, then the system complexity is reduced, but the distinguishing degree between normal and malicious accounts deteriorates, leading to low identification effectiveness
Solution Approach 1:
The system applies self-service by automatically determining its own threshold without external manual intervention. The threshold computation is performed autonomously by the system using its collected data, eliminating the need for operators to understand or configure complex parameters. This maintains simplicity from the user perspective while achieving high precision through sophisticated statistical analysis.
Solution Approach 2:
The system performs preliminary actions by collecting and analyzing reading speed data to establish the baseline distribution before setting the threshold. This preparatory phase of data collection and statistical analysis enables the system to determine an optimal threshold that maximizes distinction between normal and malicious accounts, rather than relying on arbitrary manual settings.
Data Source
AI summary
A method and an apparatus for identifying a malicious account are disclosed. Implementation of the method includes: using an effective reading speed of an account in a known normal account set as a training sample, and determining distribution data of the effective reading speed of the account in the normal account set; determining an upper reading speed threshold for the account in the normal account set according to the distribution data; acquiring a reading speed of a to-be-determined account; and determining, if the reading speed of the to-be-determined account exceeds the reading speed threshold, that the to-be-determined account is a malicious account. In the method, a fixed threshold set manually is no longer used, and a threshold is set on a statistics-based theoretical basis and is no longer randomly set manually; therefore, a malicious account for fake book viewing can be accurately identified.