Data-Driven Metric Correlation for Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face challenges in identifying the most valuable metrics for decision-making, as relevant information can change rapidly or remain stagnant, and they may not be aware of significant changes unless they regularly view specific metrics.
Innovation Solution
A computer-implemented method generates a model for each metric indicator using its attributes as independent variables, selects attributes with the greatest entropy change, and creates a graphical display of a subset of metric indicators with statistically significant changes, providing personalized and relevant information to users through a user interface.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If users regularly monitor specific metrics, then they can detect changes in those metrics, but they may miss significant changes in other metrics they do not monitor
Solution Approach 1:
The system automatically identifies and presents the most relevant metric changes to users without requiring them to manually select or monitor specific metrics. The system serves itself by using entropy calculations and statistical analysis to determine which metrics warrant user attention, eliminating the need for users to continuously evaluate which metrics to watch.
Solution Approach 2:
The system provides feedback to users about metric changes that are statistically significant and informationally relevant. By calculating entropy changes and comparing current metric values against forecasted values, the system feeds back only the most important changes to users, ensuring they receive critical information without being overwhelmed by all possible metric variations.
2Loss of information
If the system provides all metric indicators to users, then users have complete information, but the information becomes overwhelming and difficult to prioritize
Solution Approach 1:
The system extracts only the most informationally relevant metric changes from the complete set of available metrics. By calculating entropy changes and statistical significance for each metric, the system separates the critical few metrics that warrant user attention from the many less important metrics, presenting only the extracted relevant information to users.
Solution Approach 2:
The system applies different levels of information presentation to different metrics based on their individual entropy changes and statistical significance. Rather than treating all metrics uniformly, the system provides detailed information for metrics with high entropy changes while minimizing or omitting information for metrics with low entropy changes, creating a locally optimized information presentation for each metric.
3Measurement precision
If users manually analyze metric changes to identify significant ones, then they can make informed decisions, but this process consumes significant time and resources
Solution Approach 1:
The system performs preliminary analysis of all metric changes using entropy calculations and statistical tests before presenting information to users. By pre-calculating which metrics have statistically significant changes and high informational content, the system eliminates the need for users to perform time-consuming manual analysis, having already done the rigorous statistical evaluation in advance.
Solution Approach 2:
The system replaces manual human analysis of metric changes with automated computational methods including entropy calculations, time series forecasting, and statistical significance testing. This substitution of mechanical/computational analysis for human analysis maintains high measurement precision while dramatically reducing the time and cognitive resources required.
Data Source
AI summary
Described herein are techniques for identifying highly relevant content for a user to view in the form of KPI cards and providing the relevant view to the user automatically or by suggestion. The KPIs of highest practical and statistical significance are provided when the user accesses the user interface. In some embodiments, when the user is viewing a KPI, other relevant KPIs may be provided for the user to view as suggestions. Further, in some embodiments, the user may be provided with the KPIs of significance based on anomaly detection, and the explanation for the anomaly as well as suggestions for remedying any issues may be provided to the user. The highly informational content can be surfaced through the use of the advanced machine learning algorithms described herein.