用于训练异常检测算法的电信采集网络特征数据集与SOFI开源的KPI数据集
收藏资源简介:
数据集内容:该数据集主要面向网络故障的研究,基于提高用户网络使用体验、提高网络运营服务质量的需求建设。数据集包括网络的特征数据与kpi数据两部分。特征数据来源于电信网络,通过对现实用户使用网络过程中产生的网络特征数据收集而来。特征数据记录了用户特征,包括了5905条用于训练模型的数据与1478条用于测试模型的数据。kpi数据来源于开源数据集SOFI,通过对企业用网模拟收集而来,包括了用于训练、测试的数据,内含一个数据说明文件与一个三方测试报告。 采集方案:通过python运行模型,根据用户使用网络中所产生的网络数据进行分类。 采集地点:香港中文大学(深圳) 采集时间:2023年3月 设备情况:仿真平台为台式电脑 数据集包含两个文件夹、一个数据集说明文件.docx和一个三方测试报告,文件夹中为数据内容,文件夹名称分别为feature_dataset和kpi_dataset,每个文件夹中均包含两个csv文件。容量约为:2.84MB。
Dataset Content: This dataset is developed for network fault research, aiming to improve user network experience and the quality of network operation services. It consists of two parts: network feature data and KPI data. The network feature data is sourced from telecom networks, collected from network feature records generated during real users' network usage. It records user characteristics, including 5905 training samples and 1478 testing samples for model development. The KPI data is derived from the open-source dataset SOFI, collected via enterprise network simulation, containing data for both model training and testing. It includes a data description document and a third-party test report. Collection Method: Network data generated during user network usage is classified by running models through Python. Collection Location: The Chinese University of Hong Kong, Shenzhen Collection Time: March 2023 Equipment Situation: The simulation platform is a desktop computer. Dataset Package Details: The dataset contains two folders, a dataset description document (.docx) and a third-party test report. The two folders are named feature_dataset and kpi_dataset respectively, and each folder includes two CSV files. The total capacity of the dataset is approximately 2.84 MB.




