Automated User Account Classification via Identity Keys
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual classification of user accounts as human or service accounts in computer networks is labor-intensive and inefficient, often missing undiscovered service accounts, requiring repeated efforts as new accounts are added.
Innovation Solution
A machine-learning-based system that uses keys from an identity management system to create a statistical model for automatically classifying user accounts, leveraging Bernoulli Naïve Bayes modeling to map keys to probabilities, thereby enabling automated classification of unclassified accounts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual classification of user accounts is performed by analysts reading organization unit keys, then classification accuracy can be maintained through human judgment, but the process becomes labor-intensive and time-consuming
Solution Approach 1:
The patent replaces the mechanical manual process of analysts reading and interpreting organization unit keys with an automated machine-learning-based system. The system uses Bernoulli Naïve Bayes modeling to automatically classify accounts by analyzing identity management keys, eliminating the need for manual human effort while maintaining classification accuracy through probabilistic inference.
2Reliability
If manual classification methods are used, then existing accounts can be classified with human expertise, but new accounts require repeated manual effort and potentially miss undiscovered service accounts
Solution Approach 1:
The patent implements a self-service automated classification system that continuously processes both existing and new accounts without requiring repeated manual intervention. The machine-learning model automatically analyzes identity management keys and classifies accounts probabilistically, ensuring consistent application of classification criteria and eliminating human error while increasing processing throughput.
3Productivity
If automated classification is implemented using machine-learning models, then manual effort is reduced and processing speed increases, but the system requires training data and model construction complexity
Solution Approach 1:
The patent transforms the classification problem into a probabilistic parameter estimation problem using Bernoulli Naïve Bayes modeling. By changing the approach from deterministic manual classification to probabilistic automated inference based on identity management keys, the system achieves high throughput while managing complexity through statistical modeling rather than complex rule-based systems.
4Reliability
If probabilistic inference with adjustable thresholds is used, then classification reliability is enhanced and false positives reduced, but the system requires threshold tuning and validation effort
Solution Approach 1:
The patent incorporates feedback mechanisms where classification results can be reviewed and used to refine the model. The system allows for threshold adjustment based on performance metrics and feedback from manual reviews, enabling continuous improvement of classification reliability while managing deployment complexity through iterative optimization.
Data Source
AI summary
The present disclosure describes a system, method, and computer program for automatically classifying user accounts within an entity's computer network, using machine-based-learning modeling and keys from an identity management system. A system uses supervised machine learning to create a statistical model that maps individual keys or sets of keys to a probability of being associated with a first type of user account (e.g., a service account). To classify an unclassified user account, the system identifies identity management keys associated with the unclassified user account. The system creates an N-dimensional vector from the keys (where N=the number of keys), and uses the vector and the statistical model to calculate a probability that the unclassified user account is the first type of user account. In response to the probability exceeding a first threshold, the system classifies the unclassified user account as the first type of user account.