1 to 5 of 5 Results
Dec 30, 2022
Ragab, Mohamed, 2022, "Machine Fault Diagnosis", https://doi.org/10.21979/N9/PU85XN, DR-NTU (Data), V1
The dataset consists of sensor readings from bearing machines under four different operating conditions. Each of these conditions has three classes: healthy, inner-bearing damage, and outer-bearing damage. The operating conditions refer to different values for rotational speed, l... |
May 27, 2022
Ragab, Mohamed; Eldele, Emadeldeen, 2022, "UCI HAR Dataset Processed", https://doi.org/10.21979/N9/0SYHTZ, DR-NTU (Data), V1
UCIHAR is one of the most widely used datasets to evaluate performance on time series data. It contains three different sensors namely, accelerometer, gyroscope, and body sensors. These sensors have been used to collect data from 30 different persons. In our experiments, we treat... |
May 27, 2022
Ragab, Mohamed; Eldele, Emadeldeen, 2022, "Subject-wise Sleep Stage Data", https://doi.org/10.21979/N9/UD1IM9, DR-NTU (Data), V1
Sleep stage classification (SSC) problem aims to classify the electroencephalography (EEG) signals into five stages i.e. Wake (W), Non-Rapid Eye Movement stages (N1, N2, N3), and Rapid Eye Movement (REM). We adopted Sleep-EDF dataset (Goldberger et al., 2000), which contains EEG... |
May 27, 2022
Ragab, Mohamed; Eldele, Emadeldeen, 2022, "WISDM Dataset Processed", https://doi.org/10.21979/N9/KJWE5B, DR-NTU (Data), V1, UNF:6:UNCqR5C0NzMZQm8xrjVq0Q== [fileUNF]
WISDM is another popular activity recognition dataset for the evaluation of time series domain adaptation. In this dataset, accelerometer sensors were applied to collect data from 36 subjects. This data can be more challenging because of the class imbalance issue among different... |
May 27, 2022
Ragab, Mohamed; Eldele, Emadeldeen, 2022, "HHAR Processed Data", https://doi.org/10.21979/N9/OWDFXO, DR-NTU (Data), V1
The Heterogeneity Human Activity Recognition (HHAR) dataset has been collected from 9 different users using sensor readings from smartphones and smartwatches. In our experiments, we consider each user as a domain. We constructed 10 cross-domain scenarios from randomly selected us... |