Synthetic S-shaped Latent Space Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/michaelzhang01/bayesian-rflvm
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一个由具备RBF协方差核的GPLVM生成的综合数据集，旨在评估RFLVM估算S形潜在空间的能力。该数据集用于计算测试集观测值与预测观测值之间的均方误差（MSE），其中20%的观测值被保留用于测试。数据集的规模为N=500，J=100，D=2，任务是非线性降维。
This is a synthetic dataset generated by the Gaussian Process Latent Variable Model (GPLVM) with an RBF covariance kernel, intended to assess the capability of RFLVM in estimating the sigmoidal latent space. This dataset is used to calculate the mean squared error (MSE) between test set observations and predicted observations, with 20% of all observations reserved for testing. The dataset has N=500, J=100, and D=2, and the corresponding task is nonlinear dimensionality reduction.
提供机构：
Authors of the paper
