netop/gotsf-ds

Beam-Level (5G)时间序列数据集是一个专门为支持通信网络中关键性能指标（KPIs）准确预测研究而设计的多变量时间序列数据集。数据集包含四个主要配置：下行物理资源块（DLPRB）时间序列数据、下行吞吐量体积（DLThpVol）时间序列数据、下行吞吐量时间（DLThpTime）时间序列数据和测量报告数量（MR_number）时间序列数据。数据集涵盖了2,880个波束，跨越30个基站（每个基站3个小区，每个小区32个波束），时间跨度为5周训练数据和2周测试数据（第6周和第11周）。数据格式为CSV文件，包含时间列和多个波束列。该数据集旨在为网络管理和资源分配效率优化提供支持，并为时间序列模型的研究提供基准。

The Beam-Level (5G) Time-Series Dataset is a novel multivariate time series dataset specifically curated to support research in enabling accurate prediction of Key Performance Indicators (KPIs) across communication networks. The dataset includes four main configurations: Downlink Physical Resource Block (DLPRB) time-series data, Downlink Throughput Volume (DLThpVol) time-series data, Downlink Throughput Time (DLThpTime) time-series data, and Measurement Report Number (MR_number) time-series data. It comprises 2,880 beams across 30 Base Stations (3 Cells per Station, 32 Beams per Cell), with a duration of 5 weeks of training data and 2 weeks of test data (Week 6 and Week 11). The data is stored in CSV files, each containing a Time column and multiple beam columns. This dataset is designed to optimize network management and enhance resource allocation efficiency, serving as a benchmark for state-of-the-art time series models.