Stratified User Equipment Sampling for Unbiased Analytics

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing 3GPP networks lack a-priori knowledge about the distribution of user equipment (UE) populations, leading to non-representative sampling and biased analytics due to the use of simple random sampling techniques, which can result in skewed or incorrect statistical and analytic results.

Innovation Solution

Introduce a partition criteria attribute for event reporting information to enable stratified sampling by grouping UEs based on characteristics such as Type Allocation Code (TAC), Application ID, and UE communication information, ensuring that each sub-population receives a proportional sampling ratio, thereby selecting a representative subset.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If simple random sampling is used to collect UE data, then the data collection process is simple and fast, but the sampling results are biased and non-representative of the actual UE population distribution

Engineering Contradiction:
Improverepresentativeness of samplingVSAvoidcomplexity of sampling process
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies segmentation by dividing the UE population into distinct sub-populations or strata based on specific criteria (e.g., device type, service type, mobility pattern). This segmentation allows each subgroup to be sampled separately, ensuring that the overall sample accurately reflects the heterogeneous composition of the total UE population, thereby resolving the bias inherent in simple random sampling.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the sampling parameter from uniform random selection to stratified selection with proportional allocation. By introducing stratification variables and adjusting the sampling methodology to account for population distribution characteristics, the system transforms the sampling process to achieve representativeness while managing complexity through structured parameter definition.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If stratified sampling with partition criteria is implemented, then representative UE subsets are obtained, but the complexity of data collection and processing increases

Engineering Contradiction:
Improveaccuracy of analytics resultsVSAvoidcomplexity of sampling implementation
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by pre-defining partition criteria and strata classifications before the actual data collection process. The network function establishes the stratification framework in advance, including the definition of sub-population categories and their corresponding characteristics. This preliminary setup enables the sampling process to proceed systematically without requiring complex real-time decisions, thereby reducing implementation complexity while maintaining accuracy.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3952344B1Systems and methods to enable representative user equipment sampling for user equipment-related analytics services
Publication Date: 2025.09.17 NOKIA SOLUTIONS & NETWORKS OY
  • EP3952344B1 patent drawingFigure 1
  • EP3952344B1 patent drawingFigure 2
  • EP3952344B1 patent drawingFigure 3

AI summary

Systems, methods, apparatuses, and computer program products that enable representative user equipment (UE) sampling for UE-related analytics services are provided.