Network Quality Diagnosis via User Service Data Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network quality diagnosis methods fail to detect local network issues effectively, as they rely on average measurements of Key Performance Indicators (KPIs) and Key Quality Indicators (KQIs, which can mask non-important KPI deficiencies, leading to undetected problems.
Innovation Solution
A method that collects service data from all users, identifies poor-quality records through commonality analysis, and diagnoses network issues from a microscopic user service experience perspective, using techniques like inflection point analysis and preset threshold filtering to prevent local problems from being overlooked.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If average measurement of KPIs and KQIs is used to evaluate network performance, then overall network status can be effectively monitored, but local network problems are covered up and cannot be effectively located
Solution Approach 1:
The patent segments the service data from all users into multiple groups based on different dimensions (user identifier, service data generation time, server address, network address, access point name, cell identity, longitude and latitude information). By performing commonality analysis on poor-quality records within each group separately, the method can identify local network problems that would be masked in aggregate measurements, thus resolving the contradiction between overall monitoring and local detection.
2Reliability
If all service data of all users is collected and analyzed, then comprehensive network quality detection is achieved, but system complexity increases
Solution Approach 1:
The patent divides the large volume of service data into multiple groups based on different dimensions such as user identifier, time, server address, network address, access point name, cell identity, and geographic information. This segmentation reduces the complexity of analyzing all data at once while maintaining comprehensive coverage for reliable network quality diagnosis.
Solution Approach 2:
The patent extracts poor-quality records from the service data by comparing service experience values against preset thresholds. This extraction process isolates the problematic data points that need attention, allowing the system to focus analysis on relevant cases rather than processing every record, thus reducing overall system complexity while improving diagnostic reliability.
3Measurement precision
If service data is grouped by multiple dimensions for analysis, then finer granularity detection is achieved, but processing time increases
Solution Approach 1:
The patent segments service data into groups based on multiple dimensions including user identifier, service data generation time, server address, network address, access point name, cell identity, and longitude and latitude information. This multi-dimensional segmentation enables precise problem location by identifying which specific group contains the poor-quality records, thus achieving fine-grained detection without requiring exhaustive analysis of all data.
Solution Approach 2:
The patent extracts only the poor-quality records from each group by comparing service experience values against preset thresholds. This extraction approach allows the system to focus processing efforts on problematic data points rather than analyzing every record in detail, thereby reducing processing time while maintaining high measurement precision for problem location.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A method and an apparatus for implementing network poor-quality problem diagnosis are provided. The method includes: obtaining service data of all users, where the service data includes a user service experience value; identifying poor-quality records based on the service data of all the users, where the poor-quality records are service data representing that the user service experience value is deteriorated; obtaining a common characteristic of the poor-quality records by performing commonality analysis on the poor-quality records; and diagnosing, based on the common characteristic of the poor-quality records, a network poor-quality problem that causes user service experience deterioration. The present invention can prevent a local network problem from being covered up.