15 real-world datasets from Eyuboglu et al. [2022]

该数据集包含了一个预训练的分类器以及一系列数据点，每个数据点都具备以下特征：激活度、真实标签、伪标签和模式。其中，“模式”特征指出了数据中识别出的故障模式。此外，每个数据点还包含一个512维的嵌入向量、真实的类别标签以及一个用于分类的概率向量。该数据集分为低、中、高信噪比（SNR）三个类别，共包含15个数据集，其任务是进行故障模式的识别。

This dataset includes a pre-trained classifier and a series of data points. Each data point has the following features: activation value, ground-truth label, pseudo-label, and pattern. The "pattern" feature indicates the identified fault pattern in the corresponding data. Additionally, each data point also includes a 512-dimensional embedding vector, the true category label, and a classification probability vector. This dataset is divided into three categories based on signal-to-noise ratio (SNR): low, medium, and high, and it contains a total of 15 datasets. The task of this dataset is fault pattern recognition.