Wi-Fi Network Health Monitoring with ML Root-Cause Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network health monitoring in Wi-Fi networks lacks the ability to accurately predict the underlying cause of issues due to the complexity of multiple affecting factors and the scale of the network, making it difficult to provide useful insights or effective remediation.
Innovation Solution
A system utilizing supervised machine learning to identify features indicating network health, perform feature goodness classification, compute cumulative scores, and implement a correlation model to determine correlation criteria and remediation actions, combined with reinforcement learning for optimal remediation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional network monitoring methods are used to track individual parameters, then parameter values can be monitored for SLAs, but useful insight into the cause of network issues cannot be provided due to multiple affecting factors and network scale
Solution Approach 1:
The patent segments the complex network monitoring task into distinct functional modules: a problem detection model that identifies network problems using supervised machine learning, a correlation model that determines root causes by analyzing relationships between network parameters, and a remediation model that implements corrective actions. This segmentation allows each module to specialize in a specific aspect of problem-solving, improving accuracy while managing complexity through modular design
Solution Approach 2:
The patent introduces an intermediary correlation model that acts as a bridge between raw network parameters and problem diagnosis. This correlation model analyzes relationships between multiple parameters simultaneously, translating complex multi-factor interactions into actionable insights about root causes, thereby improving diagnostic accuracy without requiring direct complex analysis of all parameters
2Measurement precision
If individual parameters are monitored independently, then parameter values can be tracked, but accurate prediction of underlying network causes is difficult due to commonality of factors affecting different issues
Solution Approach 1:
The patent merges multiple individual parameter monitoring functions into a unified problem detection model that evaluates network health holistically. By combining parameters and their relationships into a single correlated analysis framework, the system preserves information about how factors interact while improving root cause prediction accuracy, avoiding the information loss that occurs when parameters are analyzed in isolation
3Extent of automation
If manual analysis of network parameters is performed, then detailed inspection is possible, but automation of problem identification and remediation is reduced
Solution Approach 1:
The patent implements preliminary action by pre-training machine learning models with historical network data and known problem patterns. The problem detection model, correlation model, and remediation model are prepared in advance to automatically identify and respond to network issues without manual intervention, significantly reducing problem detection time while maintaining high accuracy through pre-learned knowledge
Data Source
AI summary
A computer-implemented method includes identifying features including access point (AP) parameters and client device parameters that indicate a network health of a network of one or more access points (AP) and client devices, performing feature goodness classification (FGC) to classify each identified feature independently, computing cumulative scores periodically as a weighted sum of normalized values of each feature, determining if a network problem is identified based on the cumulative scores, and implementing a correlation model with supervised machine learning to determine correlation criteria and remediation actions based on features contributing to an identified network problem. The computer-implemented method may also implement a reinforcement learning (RL) based remediation model to rank intersection regions for correlation criteria based on rewards.