YD/T 4521-2023
Telecom data specification data model for machine learning (English Version)

Standard No.

YD/T 4521-2023

Language

Chinese, Available in English version

Release Date

2023

Published By

Professional Standard - Post and Telecommunication  CN  /  YD

Latest

YD/T 4521-2023

Introduction

Analysis of the Standard Core Framework

Layer	Data Type	Modeling Principles	Typical Applications
User Experience Layer	Customer Business Data/Customer Information	Business Feature Abstraction	QoE Optimization
Business Network Layer	Network Business Data	Service Decoupling	Business Orchestration
Physical Network Layer	Resource Data/Site Data	Entity Mapping	Resource Scheduling

---

Key Technical Features

Layered Decoupled Architecture Practice

The standard adopts a five-layer vertical decoupling design: from the spatiotemporal base layer (GIS coordinates, environmental data) to the user experience layer (QoE data), each layer realizes data interaction through standard interfaces such as AIPhysicalSite. In a 5G network optimization project in a certain province, China Mobile improved AI training efficiency by 40% by separating physical resource data from logical topology data.

Innovation in spatiotemporal data modeling

Innovative definition of a combined model of AITemporal time objects and AIGeoObject space objects:

• Time dimension: supports point-type time (such as alarm trigger time) and interval-type time (such as service session cycle)
• Spatial dimension: includes three levels of accuracy: point (base station location), line (transmission route), and surface (coverage area)

---

Implementation Recommendations

Data Governance Path

1. Metadata standardization: unify performance indicator definitions according to the AIDefIndicator specification
2. Relationship Construction: Use the AILink topology model to establish resource-business mapping
3. Quality Control: Establish a data credibility assessment mechanism based on the measurement data model in Chapter 5.11

Machine Learning Adaptation Solution

Algorithm Type	Applicable Data Model	Preprocessing Requirements
Time Series Forecasting	AIMetricsMeasurement Data	Time Alignment Processing
Graph Reasoning	AIKGRelationShipKnowledge Base	Relationship Weight Annotation

YD/T 4521-2023 Referenced Document

• GB/T 5271.1-2000 Information technology-Vocabulary Part1:Fundamental terms

YD/T 4521-2023 history

• 2023 YD/T 4521-2023 Telecom data specification data model for machine learning

Topics on standards and specifications

Standard and Specification