Network Data Analytics Filtering for Abnormal Device States
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network data analytics outputs in wireless communication networks can be inaccurate due to incorrect sample data, leading to suboptimal network policies and operations.
Innovation Solution
A second data analytics network element receives a status analytics output from a first element, identifies abnormal states in target objects, and generates corrected analytics outputs by excluding data from these objects, sending appropriate indication information to adjust or disable previous outputs, thereby improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If network data analytics function uses all available sample data for training and analysis, then the quantity of data increases, but the accuracy of analytics output decreases when incorrect sample data is present
Solution Approach 1:
The patent extracts and removes incorrect sample data from the dataset before performing analytics. The NWDAF identifies and excludes data from network devices in abnormal states, ensuring that only correct sample data is used for training and analysis, thereby maintaining high accuracy while using sufficient data quantity.
Solution Approach 2:
The patent changes the parameter of data quality by filtering out incorrect samples. By monitoring the operational status of network devices and excluding those in abnormal states, the system transforms the dataset from containing mixed quality data to containing only high-quality data, resolving the contradiction between data quantity and accuracy.
2Measurement precision
If network data analytics function excludes data from abnormal devices, then the accuracy of analytics output improves, but the quantity of usable data decreases
Solution Approach 1:
The patent selectively extracts only the necessary portion of data (correct samples) while discarding harmful data (incorrect samples from abnormal devices). This approach maintains adequate data quantity for effective analytics while ensuring high accuracy by excluding problematic data points.
Solution Approach 2:
The patent changes the quality parameter of the dataset by filtering based on device operational status. This transformation ensures that the remaining data, while potentially reduced in quantity, maintains high accuracy suitable for reliable analytics output.
3Productivity
If network elements generate analytics outputs without verifying data correctness, then the processing speed increases, but the reliability of analytics output decreases
Solution Approach 1:
The patent performs preliminary verification of data correctness by monitoring the operational status of network devices before analytics processing. By pre-identifying and excluding data from abnormal devices, the system ensures reliable analytics output without significantly impacting processing speed, as the filtering occurs as a preliminary step.
Solution Approach 2:
The patent implements a feedback mechanism where the NWDAF receives status information about network devices and uses this feedback to adjust the dataset. This continuous feedback loop ensures that only data from devices in normal operational state is used, maintaining high reliability while keeping the processing efficient through automated status monitoring.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
This application discloses a communication method, apparatus, and system. The communication method includes: A second data analytics network element receives a status analytics output of a target object from a first data analytics network element, where the target object includes one or more of a network device, a sub-domain of a network, an all-domain of a network, or a terminal device. The second data analytics network element obtains, based on the status analytics output of the target object, first input data corresponding to a target type of analytics, where when the status analytics output of the target object indicates that the target object is in an abnormal state, the first input data does not include data corresponding to the target object. The second data analytics network element generates, based on the first input data, a first analytics output corresponding to the target type of analytics. When the status analytics output of the target object represents that the target object is in the abnormal state, the first input data does not include data corresponding to the target object. In this way, the first analytics output may not be affected by incorrect data corresponding to the target object, so that correctness of the first analytics output can be improved.