SHAPES

该数据集是一个用于概念验证实验的低维数据集，它展示了将两个连续协变量与一个分类类别标签结合的方法。边界并不总是清晰，这表明通过增加树的数量或扩大训练集，还有改进的空间。该数据集包含了2000个样本，其任务是进行密度估计和生成建模。This is a low-dimensional dataset designed for proof-of-concept experiments, which demonstrates the method of combining two continuous covariates with a categorical class label. The decision boundaries are not always clear, indicating that there is room for improvement by increasing the number of trees or expanding the training set. This dataset contains 2000 samples, and its target tasks are density estimation and generative modeling.