Telecom KPI Prediction via Counter Correlation Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In telecommunication networks, predicting Key Performance Indicators (KPIs) is challenging due to interconnected counters and lack of standard techniques for complete KPI assessment, making it difficult for Subject-Matter-Experts to isolate issues and anticipate the impact of changes on network performance.
Innovation Solution
A system and method using data analysis techniques, such as correlation and regression analysis, to monitor and predict KPIs by identifying influencing counters and building correlation equations, enabling more accurate forecasting and qualitative insights into KPI trends.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data analysis techniques are applied to identify influencing counters and build correlation equations, then prediction accuracy of KPIs is improved, but system complexity and computational requirements increase
Solution Approach 1:
The system segments the complex task of KPI prediction by identifying and analyzing individual counters separately. The method breaks down the inter-connected counter relationships into manageable units by monitoring each counter independently and then combining their individual impacts through correlation equations, making the overall system more tractable despite the complexity of inter-counter dependencies
Solution Approach 2:
The patent introduces correlation equations as intermediary mathematical models that mediate between raw counter data and KPI predictions. These equations serve as intermediaries that capture the relationships between counters and KPIs, allowing the system to handle complexity by working with simplified mathematical representations rather than directly managing the full complexity of counter interactions
2Ease of operation
If aggregated view of performance counters is used, then ease of monitoring is improved, but ability to detect root cause is worsened
Solution Approach 1:
The system applies local quality by providing different levels of detail for different monitoring needs. While maintaining an aggregated view for easy monitoring, the patent enables drill-down capability to examine individual counter behaviors and their specific impacts on KPIs. This allows operators to access detailed local information about specific counters when root cause analysis is needed, while preserving the simplicity of aggregated views for routine monitoring
3Reliability
If standard techniques for complete KPI assessment are implemented, then reliability of KPI evaluation is improved, but device complexity increases
Solution Approach 1:
The system performs preliminary actions by pre-identifying and monitoring the specific counters that influence each KPI before actual KPI assessment is needed. The method establishes correlation equations in advance based on historical counter data, so that when KPI evaluation is required, the system can reliably assess KPIs using pre-computed relationships rather than having to analyze all possible counter interactions from scratch, thereby improving reliability without proportionally increasing complexity
Data Source
AI summary
The present disclosure relates to system(s) and method(s) for predicting a Key Performance Indicator (KPI) in a telecommunication network is illustrated. The system is configured to monitor a set of counters and a Key Performance Indicator corresponding to a telecommunication network. The set of counters and the Key Performance Indicator (KPI) are monitored for a predefined time interval to gather sample data. The system is configured to analyze the sample data using a data analysis technique in order to identify a subset of counters, from the set of counters, influencing the KPI and a correlation coefficient associated with each counter from the subset of counters, wherein the correlation coefficient associated with each counter is identified after normalizing the subset of counters. The system is configured to apply regression on the subset of counters and the KPI in order to build a correlation equation between the subset of counters and the KPI.