Synthetic Causal Direction Datasets

该数据集包含了1000个生成的数据集，每个数据集都由不同因果机制和噪声类型下的随机变量对（X, Y）组成。这些数据集涵盖了离散和连续随机变量，并具有各种噪声分布。每个数据集都用于测试所提出的方法，以从观测数据中预测因果方向。规模上，这些数据集共有1000组，每组1000对。任务目标是利用这些数据集进行从观测数据中预测因果方向的预测任务。

This dataset includes 1000 generated datasets, each of which comprises random variable pairs (X, Y) under distinct causal mechanisms and noise types. These datasets cover both discrete and continuous random variables and feature a wide range of noise distributions. Each generated dataset is designed to evaluate the proposed method for predicting causal directions from observational data. In terms of scale, each of these 1000 datasets contains 1000 variable pairs. The task objective is to utilize these datasets to perform the task of predicting causal directions from observational data.