Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
230
datasets available to search
ShareScore release 0.7.1
Dataset results
230 results for “time series data”
IRIS preprocessed data used in paper "Multi variables time series information bottleneck"
<p>Prprocessed data used in paper "Multi variables time series information bottleneck" with the <a href="https://github.com/DenisUllmann/IB-MTS">GitHub</a> code</p> <p>This dataset is created from a public available dataset of observations performed by IRIS, a NASA small explorer mission developed and operated by LMSAL with mission operations executed at NASA Ames Research Center and major contributions to downlink communications funded by ESA and the Norwegian Space Centre.</p> <p>Multiple Time Series of IRIS level 2 data are available <a href="https://iris.lmsal.com/search/">here</a></p> <p>The selected data was labeled using these definitions:</p> <p>QS: Quiet Sun<br> AR: Active Regions of the Sun<br> FL: Flare</p> <p>A time series is labeled QS when every single time step refer to a quiet sun activity.<br> When a given time series is partially composed of flaring events, the global time series is labeled as FL.</p> <p>The npz file is a numpy (np) compressed data and can be loaded using np.load with allow_pickle=True<br> Loaded data is then a python dict described bellow.</p> <p>Each sample 'data' is a np.ndarray with 2 dimensions: time (various length) and wavelength (length=240 representing a range between 2793.8401Å and 2806.02Å).</p> <p>Each sample is given a 'position' which is a list of length 4:<br> position[1] is a string that gives the name of the event<br> position[4] is a boolean vector that gives the time positionsof the corresponding sample in the original sequence of public IRIS level2 data</p> <p>Data file info :</p> <p>Type: .npz<br> Size: 11.89GB</p> <p>*** Key: 'data_TR_QS'<br> ndarray data of length 2467<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'data_TR_AR'<br> ndarray data of length 1042<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'data_TR_FL'<br> ndarray data of length 1055<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'data_VAL_QS'<br> ndarray data of length 325<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'data_VAL_AR'<br> ndarray data of length 1042<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'data_VAL_FL'<br> ndarray data of length 714<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'data_TE_QS'<br> ndarray data of length 1428<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'data_TE_AR'<br> ndarray data of length 792<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'data_TE_FL'<br> ndarray data of length 356<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'data_TR'<br> ndarray data of length 4564<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'data_VAL'<br> ndarray data of length 2081<br> containing np.ndarray of shapes ['various', 240]</p> <p><br> *** Key: 'data_TE'<br> ndarray data of length 2576<br> containing np.ndarray of shapes ['various', 240]</p> <p> </p> <p>*** Key: 'position_TR_QS'<br> ndarray data of length 2467<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_TR_AR'<br> ndarray data of length 1042<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_TR_FL'<br> ndarray data of length 1055<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_VAL_QS'<br> ndarray data of length 325<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_VAL_AR'<br> ndarray data of length 1042<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_VAL_FL'<br> ndarray data of length 714<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_TE_QS'<br> ndarray data of length 1428<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_TE_AR'<br> ndarray data of length 792<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_TE_FL'<br> ndarray data of length 356<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_TR'<br> ndarray data of length 4564<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_VAL'<br> ndarray data of length 2081<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p> <p> </p> <p>*** Key: 'position_TE'<br> ndarray data of length 2576<br> containing ndarray data of length 4<br> containing mix of types {'ndarray', 'int', 'str'}</p>
Time Series Data of Gaze, Head Pose, Hand Pose, and Object Positions for Object Approaches with a Given Intention
<p>This data set comprises time series data of gaze, head pose, hand pose, and object positions for object approaches with a given intention. The data was captured in the context of the following publication:</p> <ul> <li><em>Michael Fennel, Serge Garbay, Antonio Zea, Uwe D. Hanebeck</em>, <strong>Intention Estimation with Recurrent Neural Networks for Mixed Reality Environments</strong>, Proceedings of the 26th International Conference on Information Fusion (Fusion 2023) <em>(under review)</em></li> </ul> <p>A Microsoft Hololens 2 was used for recording the data at 60 fps under the modalities explained in detail in the above-mentioned paper.</p> <p>The file names are structured as follows:</p> <ul> <li><em>1st/2nd:</em> <ul> <li>The data with "1st" contains approaches to randomly placed objects on a grid, which are rendered in augmented reality. The user is informed about the object to approach using a visual cue. This corresponds to Section IV-A.</li> <li>The data with "2nd" contains approaches to real objects placed statically in a room. The user is informed about the object to approach using a voice command.</li> </ul> </li> <li><em>unfiltered:</em> Contains all approaches, including those where the user disrespects the given commands. Filtering is done as described in the paper.</li> <li><em>train/val/test:</em> The first dataset was split in a 70/20/10 ratio for training, validation, and test.</li> </ul> <p>Each data set contains the following columns. In each approach, 5 objects numbered from i=0 to i=4 are present.</p> <ul> <li>General: <ul> <li><em>time:</em> in seconds</li> <li><em>subject:</em> consecutive subject number</li> <li><em>handedness:</em> left (1), right (0)</li> <li><em>trial:</em> consecutive trial number per subject</li> <li><em>target_label:</em> index of the object to approach (0 to 4)</li> </ul> </li> <li>Data in world coordinates: <ul> <li><em>head_{x,y,z}:</em> head position</li> <li><em>head_quat_{w,x,y,z}:</em> head orientation quaternion</li> <li><em>W_gaze_{x,y,z}:</em> gaze direction</li> <li><em>W_r_hand_{x,y,z}:</em> right hand position</li> <li><em>W_r_hand_quat_{w,x,y,z}:</em> right hand orientation quaternion</li> <li><em>W_l_hand_{x,y,z}:</em> left hand position</li> <li><em>W_l_hand_quat_{w,x,y,z}:</em> left hand orientation quaternion</li> <li><em>W_object_i_{x,y,z}:</em> position of object i</li> <li><em>W_object_i_quat {w,x,y,z}</em>: orientation quaternion of object i</li> </ul> </li> <li>Data in egocentric coordinates (head coordinate system). This data is provided for convenience and can be derived from the other data: <ul> <li><em>gaze_{x,y,z}:</em> gaze direction</li> <li><em>r_hand_{x,y,z}:</em> right hand position</li> <li><em>r_hand_quat_{w,x,y,z}:</em> right hand orientation quaternion</li> <li><em>l_hand_{x,y,z}:</em> left hand position</li> <li><em>l_hand_quat_{w,x,y,z}:</em> left hand orientation quaternion</li> <li><em>object_i_{x,y,z}:</em> position of object i</li> <li><em>object_i_quat {w,x,y,z}</em>: orientation quaternion of object i</li> </ul> </li> </ul> <p><strong>Acknowledgment:</strong></p> <p>This work was supported by the <a href="https://robdekon.de/">ROBDEKON</a> project of the German Federal Ministry of Education and Research.</p>
LAGOS-NE v.1.054.1 Lake water clarity time series (1987-2011), climate, and geophysical data for 601 lakes across a 17-state region of the United States
Time series of median summer water clarity (secchi) values from 601 unique lakes in the Midwest and Northeast United States. Water clarity observations were derived from the Lake Multi-Scaled Geospatial and Temporal Database LAGOS-NELIMNO version 1.054.1. These data were used to assess long-term changes in water clarity from 1987-2011, and the potential drivers of those trends (Lottig et al. in press). Summer open water period was used to approximate the stratified period in the study lakes, which was defined as June 15 to September 15. Over the 25-year time period, each lake had to have at least a single summer water clarity observation for 22 of 25 years. The median number of secchi measurements that were used to derive a single annual median value for each lake was approximately 9. Of the over 14,000 annual estimates of water clarity that we generated, only two percent of those annual values were generated from a single observation and median number of observations for each lake over the 25-year study period was 223. Each unique lake with water clarity data also has supporting geophysical data, including climate, land use, hydrology, and topography derived at multiple spatial scales. Lake-specific characteristics, such as depth and area, are also reported. The geospatial data came from LAGOS-NEGEO version 1.03 except for the annual climate data which was aggregated at the HUC8 spatial scale from monthly PRISM data. For more specific information on how LAGOS-NE was created, see Soranno et al. 2015. Citations: Lottig, N.R., P-N. Tan, T. Wager, K.S. Cheruvelil, P.A. Soranno, E.H. Stanley, C.E Scott, C.A. Stow, and S. Yuan. in press. Macroscale patterns of synchrony identify complex relationships among spatial and temporal ecosystem drivers. Ecosphere Soranno P.A., Bissell E.G., Cheruvelil K.S., Christel S.T., Collins S.M., Fergus C.E., Filstrup C.T., Lapierre J.-F., Lottig N.R., Oliver S.K., Scott C.E., Smith N.J., Stopyak S., Yuan S., Bremigan M.T., Downing J.A., G
Applying time series analyses on continuous accelerometry data – Dataset
<p>Data and analysis script accompanying the study:</p> <p>Applying time series analyses on continuous accelerometry data – a clinical example in older adults with and without cognitive impairment</p>
Woody Cover Mapping in the Kruger National Park using Sentinel-1 time series and LiDAR data
<p>This data repository presents a workflow to derive woody cover information for the Kruger National Park, South Africa, from freely available Sentinel-1 C-Band time series and LiDAR data (modified from Smit et al. 2016) using machine learning (MLR and Ranger in R). The methodology is described in following publication:</p> <p><em>Urban, M., K. Heckel, C. Berger, P. Schratz, I.P.J. Smit, T. Strydom, J. Baade & C. Schmullius (2020): Woody Cover Mapping in the Savanna Ecosystem of the Kruger National Park Using Sentinel-1 C-Band Time Series Data. Koedoe.</em></p> <p>In order to derive woody cover percentage information, download all files into one folder and run the R-Files consecutively from 01_ to 04_. Follow the instruction within each of the R-Files, which are written as comments in the programming code.</p> <p>The data repository consist of the following files:</p> <p><strong>R-Files:</strong></p> <p>1. Script 1: 01_MLR_tune_spatial_final</p> <p>2. Script 2: 02_MLR_cross_validation_spatial_final</p> <p>3. Script 3: 03_MLR_RANGER_train_final</p> <p>4. Script 4: 04_MLR_prediction_woody_cover_final</p> <p> </p> <p><strong>Training dataset - ENVI FILE (layerstack of Sentinel-1 VH and VV backscatter between 2016 and 2017 and the woody cover reference derived from the LiDAR data) :</strong></p> <p>1. S1_A_VH_VV_16_17_lidar</p> <p> </p> <p><strong>Data for prediction - ENVI FILES (3 example regions in the Kruger National Park):</strong></p> <p>1. S1_A_VH_VV_16_17_subset_example_Letaba_Rest_Camp</p> <p>2. S1_A_VH_VV_16_17_subset_example_Lower_Sabie</p> <p>3. S1_A_VH_VV_16_17_subset_example_Pafuri</p> <p> </p> <p><strong>Final woody cover maps of the Kruger National Park:</strong></p> <p>1. xx_woody_cover_map_final.rar (contains final maps in 10m, 30m, 50m and 100m spatial resolution as .tif and a QGIS project)</p> <p> </p> <p><em>References:</em></p> <p>Smit, I.P.J., Asner, G.P., Govender, N., Vaughn, N.R. & Wilgen, B.W. van, 2016, ‘An examination of the potential efficacy of high-intensity fires for reversing woody encroachment in savannas’, <em>Journal of Applied Ecology</em>, 53(5), 1623–1633.</p>
A Tool for Uncertainty Quantification in Reconstructing Sparse Water Quality Time Series Data to Assess Risk Metrics for Watershed Health and TMDL Analysis
<p>The uploaded file contains the input and output data which can be used to reproduce the results in the research article 'Uncertainty Quantification in Reconstruction of Sparse Water Quality Time Series: Implications for Watershed Health and Risk-Based TMDL Assessment'. Please refer to the file '<a href="https://zenodo.org/api/files/31b59cce-8eb2-4ee7-93aa-61474c6f6359/dst_2019_SJRW_TP_TDS.zip?versionId=2af2b54d-d5fb-4720-919d-de2e827595e2">dst_2019_SJRW_TP_TDS.zip'</a> for updated files..</p>
Three-dimensional GNSS Time Series Data for Terrestrial Water Storage Changes Inversion in Yunnan, China
<p>The dataset includes three-dimensional GNSS time series data featured in the publication "Using the global navigation satellite system and precipitation data to establish the propagation characteristics of meteorological and hydrological drought in Yunnan, China", published in 'Water Resources Research'.</p> <p>Reference:<br>Zhu, H., Chen, K., Hu, S., Liu, J.,Shi, H., Wei, G., et al. (2023). Using the global navigation satellite system and precipitation data to establish the propagation characteristics of meteorological and hydrological drought in Yunnan, China. Water Resources Research, 59, e2022WR033126. https:// doi.org/10.1029/2022WR033126</p> <p><br>The sitelist file lists basic information about all the utilized stations, including their names and geographic coordinates. <br>The Time.mat file contains the time vectors of the data employed. <br>The Filter_time_series_N/E/U.mat files showcase the filtered time series, which have been processed using Independent Component Analysis (ICA) for the inversion of terrestrial water storage in Yunnan, after removing the effects of outliers, steps, and non-tidal atmospheric/oceanic loading.</p>
Input GNSS time series data for Tanaka et al. (2024), JGR Solid Earth
<p>Detail explanatios are in the uploaded README file. </p>
Mapping Tree Species Fractions in Temperate Mixed Forests Using Sentinel-2 Time Series and Synthetically Mixed Training Data
<p>This dataset contains the latest version of a selection of result data of the paper "Mapping Tree Species Fractions in Temperate Mixed Forests Using Sentinel-2 Time Series and Synthetically Mixed Training Data" (DOI: https://doi.org/10.1016/j.rse.2025.114740 )</p> <p>The dataset contains:</p> <ol> <li>A geopackage of training points of pure tree species</li> <li>The resulting 12-band tree species fraction map of Rhineland-Palatinate</li> <li>HSV-colored map of dominant tree species. For information which tree species are represented by the different colors, refer to the Supplemental in the original paper.</li> <li>CSV-table of predicted and reference propotion of the tree species in the validation polygon (the original polygon data can not be published due to data privacy regulations) </li> </ol> <p> </p>
Assessing land surface phenology in Araucaria-Nothofagus forests in Chile with Landsat 8/Sentinel-2 time series - Data and Material
<p>This dataset contains the Enhanced Vegetation Index (EVI) data used in our research work about land surface phenology of Andean Araucaria-Nothofagus forests as well as the phenology information derived from it.</p> <p>Study area: Conguillío National Park, Chile<br> Study period: 2016-2020</p> <p>Description of datasets:</p> <p>conguillio.sen2.lnd8.evi.2016.2020.nc - A raster dataset (NetCDF) of EVI values (resolution 10m). EVI was calculated from Level-2 Sentinel-2 and Landsat 8 data. To ensure harmonization, the Landsat 8 data was resampled and reprojected to Sentinel-2 properties prior to the index calculation.</p> <p>evi_gb_beck_white.tif - A raster dataset (GeoTiff) of phenological metrics per year (resolution 10m). Metrics were derived by fitting a double logistic function (see Beck et al., 2006) to the smoothed and interpolated EVI pixel time series. Subsequently, the main phenological variables SOS (start of season) and EOS (end of season) were extracted using a 50% threshold value. The dataset itself is a result of the R package "greenbrown" and the layers are named accordingly (see https://greenbrown.r-forge.r-project.org/phenology.php). It is available as GeoTIFF and as R rasterfile.</p> <p>Details about the methodology and results describing this dataset can be found in the following publication:<br> Kosczor, E., Forkel, M., Hernández, J., Kinalczyk, D., Pirotti, F. & Kutchartt, E., 2022. Assessing land surface phenology in Araucaria-Nothofagus forests in Chile with Landsat 8/Sentinel-2 time series. Int. J. Appl. Earth Obs. Geoinf. 112, 102862. https://doi.org/10.1016/j.jag.2022.102862</p>
Data associated with the manuscript "Simple statistical models can be sufficient for testing hypotheses with population time series data"
<p>This is a revised version of the archive of R code and data used in the manuscript, <em>Simple statistical models can be sufficient for testing hypotheses with population time series data. </em>The data are in three files. <em>etodata1.csv</em> and <em>etodata2.csv</em> contain two versions of the same data for shoal-dwelling fishes in the Etowah River and associated environmental covariates. <em>knz_dat</em> contains data for small mammals collected in the Konza Prairie Biological Station and associated environmental covariates. The R code consists of four primary files that call nine auxiliary files. CaseStudy1-main_code and CaseStudy2-main_code are the primary files for running the two case studies. Simulations1 and Simulations2 are the files for running the two batteries of simulations. We thank the Konza Prairie Biological Station and Konza Prairie Long-Term Ecological Research Program supported by the National Science Foundation (DEB-1440484) for collecting and providing access to mammal community data. More details are in the manuscript and supporting information. </p>
Code and Data associated with "Discovery of positive and purifying selection in metagenomic time series of hypermutator microbial populations"
<p>Code and data sufficient to reproduce analyses in "Discovery of positive and purifying selection in metagenomic time series of hypermutator microbial populations".</p>
Single pixel s*t Landsat time series training data for CNN
<p>Single pixel s*t Landsat time series classification using 1D CNN</p> <p>Sep 22, 2022 update (version 2): <br> The 1D CNN classification codes are available at https://github.com/hankui/cnn_Landsat_time_series_classification_v2-Python</p> <p>The NLCD training data is available at 10.5281/zenodo.7106054</p> <p>The NLCD training data is derived from Landsat 5/7 analysis ready data (ARD) in year 2011 (as x predictor variable) and National Land Cover Database (NLCD) 2011 (as y response variable)</p> <p>The NLCD training data is distributed across Continental United States (CONUS) with 3,314,439 30m pixel locations</p> <p>The NLCD training data include (i) NLCD label with 15 classes, i.e., all NLCD classes except ice (https://www.mrlc.gov/data/legends/national-land-cover-database-class-legend-and-description)<br> (ii) year 2011 growing season Landsat ARD percentiles for Landsat 5/7 bands 2, 3, 4, 5 and 7 and for 8 band ratios derived from the five bands <br> (iii) percentiles include 10th, 20th, 25th, 30th, 35th, 40th, 50th (median), 60th, 65th, 70th, 75th, 80th, 90th so that <br> one can use 5 percentiles (10th, 25th, 50th, 75th, and 90th)<br> 7 percentiles (10th, 20th, 35th, 50th, 65th, 80th, and 90th)<br> 9 percentiles (10th, 20th, 30th, 40th, 50th, 60th, 70th, 80th, and 90th)<br> (iv) the pixel location represented in Landsat ARD tile h and v no. and the pixel i and j locations in the tile<br> (v) the no. of the cloud free observations in 2011 growing season derived for the pixel location<br> </p> <p>#*************************************************************************************************************#</p> <p>A munuscript describing how the data were derived and how the 1D CNN was adapted to the data is in review </p> <p><br> #*************************************************************************************************************#</p> <p>The codes were written in python (v3.7) and tensorflow (v2.6). </p> <p>The parameters are:</p> <p>(1) learning rate: cnn training initial learning rate 0.01 used in the paper </p> <p>(2) epoch: cnn training epochs 70 used in the paper </p> <p>(3) method: cnn training optimizer method 1: Adam method 2: dynamic learning rate used in the paper</p> <p>(4) L2: L2 regularization value; 0.001 used in paper </p> <p>(5) layer: no. of CNN layers (can be 4, 5 and 8) and 5 and 8 used in the paper</p> <p>(6) perc: training data percentages (can be 0.1, 0.5 and 0.9) tested in the paper; the evaluation is used the left 10% </p> <p>(7) gpui: which gpu process it will use (only applicable with multi-gpus) </p> <p>(8) IMG_HEIGHT: the no. of percentiles (can be 3, 5, 7 and 9) and 5, 7 and 9 used in the paper </p> <p>An example would be: </p> <p>version=7_4 </p> <p>layer=5; perc=0.1; gpui=0;IMG_HEIGHT=5</p> <p>method=0; learning_rate=0.01; epoch=10; iter=1; L2=0.001; sleep ${SLEEP}; ## Hank layer=5; perc=0.1; </p> <p>echo "python Pro_2d1d_CNN_v${version}.py ${learning_rate} ${epoch} ${method} ${L2} ${layer} ${perc} ${gpui} ${IMG_HEIGHT} "</p> <p>python Pro_2d1d_CNN_v${version}.py ${learning_rate} ${epoch} ${method} ${L2} ${layer} ${perc} ${gpui} ${IMG_HEIGHT} > layer${layer}.p${perc}.d${IMG_HEIGHT}.rate${learning_rate}.e${epoch}.L${L2}.v${version} & </p> <p><br> #*************************************************************************************************************#</p> <p>Aug 29, 2021 (version 1): <br> Training data: There are 2 input text files (csv) storing the 3,314,439 NLCD and 484,476 CDL land cover training samples:<br> NLCD training: ./NLCD/metric.ard.nlcd.Mar01.18.40.txt<br> CDL training: ./CDL/metric.ard.nlcd.Mar01.18.40.txt</p> <p>The codes and their usages are at: <br> https://github.com/hankui/cnn_Landsat_time_series_classification_v1-R<br> </p>
Station M time series study (NE Pacific) CTD data (cruises 2006-2022, surface to 4000 m depth)
<p>These datasets are from sensors mounted on remotely operated vehicle deployments (ROVs Tiburon and Doc Ricketts) to Station M (approx 4000 m) in the NE Pacific. Collection dates were from 2006 to 2022 as the ROV operated from the surface to the abyssal seafloor.</p> <p>The CTD was a Seabird SBE 21, Oxygen came from a pair of Seabird SBE 43s, Beam transmission from a Wetlabs C-Star, 25cm path, 720nm(red) </p> <div>21 and 43s were calibrated annually at Seabird, and the 43s were corrected a couple of times a year with bottle titration. </div> <div> </div> <div>Units:</div> <div> <table> <tbody> <tr> <td>depth (meters)</td> </tr> <tr> <td>heading (degrees)</td> </tr> <tr> <td>temperature (degrees C)</td> </tr> <tr> <td>salinity (unitless)</td> </tr> <tr> <td>oxygen (ml/l)</td> </tr> </tbody> </table> </div>
Data from: Evaluating Window Size Effects on Univariate Time Series Forecasting with Machine Learning
<p>In the realm of time series prediction modeling, the window size (w) is a critical hyperparameter that determines the number of time units included in each example provided to a learning model. This hyperparameter is crucial because it allows the learning model to recognize both long-term and short-term trends, as well as seasonal patterns, while reducing sensitivity to random noise. This study aims to elucidate the impact of window size on the performance of machine learning algorithms in univariate time series forecasting tasks. To achieve this, we employed 40 time series from two different domains, conducting experiments with varying window sizes using four types of machine learning algorithms: Bagging, Boosting, Stacking, and a Recurrent Neural Network (RNN) architecture. The results reveal that increasing the window size generally enhances the evaluation metric values up to a stabilization point, beyond which further increases do not significantly improve predictive accuracy. This stabilization effect was observed in both domains when w values exceeded 100 time steps. Moreover, the study found that RNN architectures do not consistently outperform ensemble models in various univariate time series forecasting scenarios.</p>
Figure 2.Flow Chart for Data preprocessing & Training-Comparative study of Financial Time Series Prediction by Artificial Neural Network with Gradient Descent Learning
<p>Methodology<br> This paper develops an ANN based comparative predictive model for NASDAQ stock<br> prediction. The first ANN model is developed with Multi-Layer Feed forward Network<br> Architecture & the second model is developed with Recurrent Neural Network Architecture. In this<br> paper gradient descent based back propagation learning algorithm is used for the supervised<br> learning of the predictive network.</p>
Data from: Using Landsat time-series to investigate nearly 50 years of tree canopy cover change across an urban-rural landscape in southern Ontario
<p><strong>Paper Abstract:</strong></p> <p>Canadian urban and adjacent landscapes have been dynamic over the last 50 years due to land management, land cover alternations, climate change, and disturbances. Remote sensing, particularly the Landsat archive, provides the only means to spatially quantify these long-term dynamics locally. Here, we explore the utility of Landsat, including the often-forgotten MSS sensor, for investigating percent tree canopy cover (TCC) change between 1972 and 2020 in a Canadian urban-rural context. We build a TCC time-series by training random forest models using visually interpreted TCC from high-resolution imagery. Predictors include topographic and yearly LandsatLinkr-harmonized and LandTrendr-fitted tasseled cap indices. Yearly binary TCC maps are built to mask consistently treeless areas and limit noise. To increase confidence in observed TCC change without historical reference imagery, we investigate multiple temporal validation options. Our TCC time-series (R2: 0.89, RMSE: 10.7%), quantifies TCC dynamics while limiting erroneous change and predictor space extrapolation. We explore TCC changes across landscapes, revealing periods of gain and loss associated with agricultural reforestation (1978-1996), housing development (on-going), drought (late 1990s), emerald ash borer (2010s), an ice storm (2013), and other drivers. Results demonstrate how long-term Landsat time-series can be used to better understand historical tree canopy change at local-regional scales. </p> <p> </p> <p><strong>Dataset details:</strong></p> <p>See paper. </p> <ul> <li>cc_72to20.tif: Yearly tree CC predictions (1972-2020)</li> <li>always_nonforest10_nowater.tif: continuous-non-canopy mask</li> <li>water.tif: water mask</li> <li>Yearly.zip: Annual predictors (including CC10) and asc outputs</li> </ul> <p> </p> <p>See code on GitHub: <a href="https://github.com/ZZMitch/PredictTreeCC_Landsat_1972to2020">ZZMitch/PredictTreeCC_Landsat_1972to2020: Code from the portion of my PhD about using Landsat time-series to predict tree canopy cover from 1972 - 2020. Code will be released as papers are published. (github.com)</a></p>
Extended data tables to Haering and Habermann, F1000Res, RNfuzzyApp: an R shiny RNA-seq data analysis app for visualisation, differential expression analysis, time-series clustering and enrichment analysis
<p><b>Background</b> </p> <p>RNA-seq is a widely adopted affordable method for large scale gene expression profiling. However, user-friendly and versatile tools for wet-lab biologists to analyse RNA-seq data beyond standard analyses such as differential expression, are rare. Especially, the analysis of time-series data is difficult for wet-lab biologists lacking advanced computational training. Furthermore, most meta-analysis tools are tailored for model organisms and not easily adaptable to other species.</p> <p><b>Results</b></p> <p>With RNfuzzyApp, we provide a user-friendly, web-based R-shiny app for differential expression analysis, as well as time-series analysis of RNA-seq data. RNfuzzyApp offers several methods for normalization and differential expression analysis of RNA-seq data, providing easy-to-use toolboxes, interactive plots and downloadable results. For time-series analysis, RNfuzzyApp presents the first web-based, automated pipeline for soft clustering with the Mfuzz R package, including methods to aid in cluster number selection, Mfuzz loop computations, cluster overlap analysis, as well as cluster enrichments.</p> <p><b>Conclusion</b></p> <p>RNfuzzyApp is an intuitive, easy to use and interactive R shiny app for RNA-seq differential expression and time-series analysis, offering a rich selection of interactive plots, providing a quick overview of raw data and generating rapid analysis results. Furthermore, its orthology assignment, enrichment analysis, as well as ID conversion functions are accessible to non-model organisms.</p>
Code and data used for the study: 'BioDeepTime: a database of biodiversity time series for modern and fossil assemblages'
<p>The repository includes code and data to reproduce the results in the manuscript ‘BioDeepTime: a database of biodiversity time series for modern and fossil assemblages' by Smith et al. (<code>analysis_biodeeptime.zip</code>).</p>
Data from: Time series of bird abundances, land cover and temperature from standardized breeding bird monitoring schemes (line transects and point count routes) from Norway, Sweden and Finland, for 1975-2016
<p><span>These data on bird species abundance and environmental variables were used in testing and comparing two different species distribution model validation methods that are applied to models which are used to predict the effects of climate change on species' distributions. The aim of the study was to investigate whether different validation methods give different results of the model's predictive performance and to demonstrate that validation methods based on measuring and validating a "static" pattern in distribution can assess model performance over-optimistically compared to methods based on measuring and validating a "change" in the distribution, which can assess the predictive performance more critically. </span></p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.