金融时序知识图谱的风险辨识与推理模型数据集

该数据集为指标4.2金融时序知识图谱的风险辨识与推理模型数据集的相关数据，具体包括模型代码、训练与测试数据集。模型代码包括了3个主要模型的源代码与相关文档，模型分别为：基于多源超图图神经网络的企业信用风险检测模型、基于逻辑规则的关键节点辨识技模型和基于元学习和多任务采样的黑天鹅事件预测模型。训练与测试数据集具体分别为：Cora 引文网络、DGraph 金融网络、5197 家企业信用风险数据集 MSGraphFin和DBLP作者网络。以及处理网络数据需要的软件工具torch_geometric。数据量大小约为5 G。

This dataset comprises the relevant data for the risk identification and reasoning model dataset of Indicator 4.2 Financial Temporal Knowledge Graph. Specifically, it includes model code, training and test datasets. The model code contains the source codes and related documentation of three core models: the Enterprise Credit Risk Detection Model based on multi-source hypergraph neural network, the Key Node Identification Model based on logical rules, and the Black Swan Event Prediction Model based on meta-learning and multi-task sampling. The training and test datasets consist of the Cora citation network, DGraph financial network, 5,197-enterprise credit risk dataset MSGraphFin, and DBLP author network. Additionally, the software tool torch_geometric required for network data processing is included. The total data volume is approximately 5 GB.

搜集汇总

背景与挑战

背景概述

该数据集为金融时序知识图谱的风险辨识与推理模型提供相关数据，包含三个主要模型的源代码及文档，以及Cora引文网络、DGraph金融网络等训练与测试数据集。数据集还附带了处理网络数据所需的软件工具，总数据量约为5 G。