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.
725
datasets available to search
ShareScore release 0.9.0
Dataset results
725 results for “Recommendation”
Making ICRP Recommendations 'Fit For Purpose' for the Response to a Nuclear or Radiological Emergency
<p><em>Background:</em> ICRP recommendations need to be made fit for purpose for the response to a nuclear or radiological emergency. While the System of Radiological Protection has performed well to meet its primary objective ‘to contribute to an appropriate level of protection for people…’ in normal, routine activities (planned exposure situations), the System has failed in its objective related to protection of the people in emergency exposure situations. This is evident from the excess deaths, increases in morbidity (e.g., diabetes, heart diseases) and other severe effects (e.g., mental, economic, social) observed after the Fukushima Daiichi nuclear power plant accident. These adverse effects resulted from protective actions taken to avert radiation exposure, consistent with international recommendations, even though the radiation health effects avoided are negligible and not discernible. These actions were therefore not justified based on the hazard to health from radiation exposure. Too much emphasis is given to protection of people from the radiation, without due consideration to the impact of the protective action. Thus justification – to do more good than harm – is not being realised. The principle of justification has not been thoroughly or correctly considered before, in part because no tools were provided to allow the risk of protective actions to be put in perspective, so our analysis will provide a technical basis for the development of such tools.</p> <p><em>Methods:</em> Meta-analysis of more than 600 papers focusing on protective actions and non-radiological health effects for different disasters (including nuclear and natural disasters). Analysis in terms of absolute risk (cases per 1000) of the health effects resulting from protective actions and radiation health effects from the averted dose based on the LNT fatal risk coefficient of 5% per Sv.</p> <p><em>Results</em>: Preliminary results are presented. We estimate approximately 15 excess early deaths per 1000 among the general population and 30 among those in long term care facilities following dislocations resulting from protective actions, while only 1 hypothetical excess radiation-induced cancer (late) death per 1000 would be averted if protective actions are taken at 20 mSv total effective dose – a criterion used in many countries for implementation of protective actions.</p> <p><em>Conclusions:</em> Implementing protective actions can result in more excess deaths than assumed to result from the radiation expose averted. Such negative impacts to the health of people must be accounted for in the System of Radiological Protection and recommendations concerning their application, or they are failing to protect people effectively.</p>
Recommender Evaluation Results and Dataset
<p><strong>Datasets:</strong></p> <ol> <li>MongoDB_Dataset: an archived MongoDB data used in this research. To reproduce/operate the proposed framework, this dataset must be restored in your local machine.</li> <li>Survey_Form: an example survey we used to evaluate our framework.</li> <li>Survey_Responses: an Excel file that contains summarized survey responses that we collected from the survey with participants. This file includes 30 responses in a raw-answer format.</li> </ol>
Mildew ratings and yields of winter wheat and spring oat varieties on NIAB and AHDB Recommended Lists, 1972-2022
<p>Data on powdery mildew ratings and yields in fungicide-treated trials relative to controls, for winter wheat and spring oat varieties on UK Recommended Lists from 1972 to 2022. These data are used in the graphs in Figure 1 of Brown & Wulff (2022) 'Diversifying the menu for crop powdery mildew resistance', Cell, DOI https://doi.org/10.1016/j.cell.2022.02.003. Data are compiled from published information. (c) NIAB for data from 1972 to 2001. (c) Agriculture and Horticulture Development Board </p>
Scientific Data Item Recommendation Benchmark Corpus
<p>Open-source generalized benchmark corpus to evaluate scientific data item (e.g. dataset or paper) recommendation methods.</p>
OpenML Dataset for Data and Algorithm Recommendations
<p>This dataset contains data obtained from the online platform OpenML [1] and represents interactions between users, algorithms, and datasets. For example, a user could apply an algorithm to a dataset for a classification task. Furthermore, it includes textual descriptions for the datasets and algorithms.</p> <p>* 512 users<br> * 12,344 algorithms<br> * 2,677 datasets<br> * 10,945 interactions</p> <p>The format of interactions.csv is (interaction id, task type, user id, dataset id, algorithm id).<br> The format of datasets.csv and algorithms.csv is (id, name, description).</p> <p>[1] https://www.openml.org/</p>
The Good First Issue Recommendation Dataset from "GFI-Bot: Automated Good First Issue Recommendation on GitHub"
<p>This is a good first issue (GFI) recommendation dataset created from the GFI-Bot project (<a href="https://github.com/osslab-pku/gfi-bot">https://github.com/osslab-pku/gfi-bot</a>). For more information about the GFI recommendation problem and GFI-Bot, please check our publications:</p> <ul> <li>Wenxin Xiao, Hao He, Weiwei Xu, Xin Tan, Jinhao Dong, and Minghui Zhou. 2022. Recommending Good First Issues in GitHub OSS Projects. In Proceedings of the 44th International Conference on Software Engineering, ICSE 2022, Pittsburgh, PA, USA, May 21–29, 2022. ACM. <a href="https://hehao98.github.io/files/2022-recgfi.pdf">https://hehao98.github.io/files/2022-recgfi.pdf</a></li> <li>Hao He, Haonan Su, Wenxin Xiao, Runzhi He, and Minghui Zhou. 2022. GFI-Bot: Automated Good First Issue Recommendation on GitHub. In Proceedings of the 2022 ACM 30th Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering, ESEC/FSE 2022, Singapore, November 14-16, 2022. ACM. <a href="https://hehao98.github.io/files/2022-gfibot.pdf">https://hehao98.github.io/files/2022-gfibot.pdf</a></li> </ul> <p>The dataset is a MongoDB dump and needs to be restored to a MongoDB instance before use. This can be done via the official <a href="https://www.mongodb.com/docs/database-tools/mongorestore/"><code>mongorestore</code></a> tool by running a command like this in the <code>dataset/</code> folder:</p> <pre><code class="language-bash">mongorestore --uri={{ your mongodb url }} --gzip </code></pre> <p>In the <code>gfibot.dataset</code> collection, each document describes the state of an issue at a certain time (either at the time of issue creation or at the time of issue resolution). The <code>resolver_commit_num</code> is the ground truth label (i.e., # of commits the issue resolver has made in the repository before issue resolution, excluding commits for resolving the issue itself; <code>resolver_commit_num = 0</code> means the resolver is someone completely new to the repository). The remaining fields can be used as features or further analyzed to derive new features.</p> <p>The <code>gfibot.resolved_issue</code> collection additionally provides information about which GitHub user resolved this issue and in what commit or pull request. This information can be used to study problems like, e.g., personalized good first issue recommendation or newcomer retention mechanisms.</p> <p>This dataset can be used to evaluate new GFI recommendation approaches. We hope it will be helpful in advancing GFI recommendation research and other future studies on open-source software onboarding.</p>
Supplementary material for 'Deployment pipeline: a review, a recommended model and its usage'
<p>This is the replication package of the paper "Deployment pipeline: a review, a recommended model and its usage".</p> <p>This replication package contains two .csv files.</p> <p>The file named SLR-initial-list.csv contains the initial list of papers retrieved from having applied a single iteration of forward snowballing using the work of Hilton et al. [1] as starting point.</p> <p> <br> The file named SLR-final-corpus.csv contains the final corpus resulting from having the study selection step as reported in the paper. This final corpus contains 32 papers.<br> </p> <p>Reference:<br> [1] Michael Hilton, Timothy Tunnell, Kai Huang, Darko Marinov, and Danny Dig. 2016. Usage, Costs, and Benefits of Continuous Integration in Open-Source Projects. In Proceedings of the 31st IEEE/ACM International Conference on Auto- mated Software Engineering (Singapore, Singapore) (ASE 2016). Association for Computing Machinery, New York, NY, USA, 426–437. https://doi.org/10.1145/ 2970276.2970358</p>
fNIRS Brain Imaging in Motor Paradigms: Evidence-Based Methodological Recommendations
<p>Anonymized behavioural, physiological, motion-capture, and <em>f</em>NIRS data and associated codes supporting the analyses reported in the paper "<em>f</em>NIRS Brain Imaging in Motor Paradigms: Evidence-Based Methodological Recommendations".</p>
Data from: Choice of prioritization method impacts recommendations for climate-informed bird conservation in the United States
<p class="MsoNormal">Climate-informed spatial planning is urgently needed to guide initiatives aimed at both conserving biodiversity as a whole (e.g., protection of 30% of lands and waters by 2030) and recovering North American avifauna in particular. Various methods for prioritizing conservation areas exist, yet alternative methods may direct managers to different lands for protection and thus varying recommendations for meeting area-based targets. Here, we used bird species distribution models and landcover projections to systematically evaluate two widely-used methods for prioritizing areas most likely to facilitate the persistence of multiple species under climate change: (1) <em>in situ </em>macrorefugia, identified as areas of high predicted species retention; and (2) complementarity-based optimizations, identified using the Zonation conservation planning software. For 17 biogeographical groups in the continental United States, we compared priority areas for bird conservation derived from these two alternatives with respect to their spatial distributions and consensus (i.e., overlap), expected conservation outcomes (e.g., species and functional diversity), predicted climate change exposure, habitat characteristics, landscape configurations, and degree of formal protection. Spatial distributions of priority areas differed by biogeographical group and method, with 40.5% consensus on average across groups. Consensus was extensive within mountainous and coastal regions and limited at high latitudes (e.g., Alaska) and in flat, interior regions (e.g., grasslands). As expected, complementarity-based optimizations more efficiently represented species than retention-based <em>in situ </em>macrorefugia, especially for forest groups, and had greater overall biodiversity value and better habitat condition. Conversely, <em>in situ </em>macrorefugia encompassed higher elevations and larger contiguous patches and were expected to experience less winter-season warming. Formal protection averaged <50% across biogeographical groups, regardless of prioritization method. Our findings illustrate the value of complementarity-based optimizations for bird conservation under climate change. More broadly, comparing approaches for prioritizing areas for long-term species persistence can reveal critical tradeoffs in recommendations for climate-informed protected area planning.</p>
Data set for the paper "Predicting Relevance of Change Recommendations"
<p>Data set for the paper Predicting Relevance of Change Recommendations by Thomas Rolfsnes, Leon Moonen, and David Binkley, In International Conference on Automated Software Engineering (ASE), pp. 694–705. 2017, IEEE.</p> <p>Please cite this work by referring to the corresponding conference publication (a preprint is included in this package).</p> <p>Abstract: Software change recommendation seeks to suggest artifacts (e.g., files or methods) that are related to changes made by a developer, and thus identifies possible omissions or next steps. While one obvious challenge for recommender systems is to produce accurate recommendations, a complimentary challenge is to rank recommendations based on their relevance. In this paper, we address this challenge for recommendation systems that are based on evolutionary coupling. Such systems use targeted association-rule mining to identify relevant patterns in a software system's change history. Traditionally, this process involves ranking artifacts using interestingness measures such as confidence and support. However, these measures often fall short when used to assess recommendation relevance. We propose the use of random forest classification models to assess recommendation relevance. This approach improves on past use of various interestingness measures by learning from previous change recommendations. We empirically evaluate our approach on fourteen open source systems and two systems from our industry partners. Furthermore, we consider complimenting two mining algorithms: CO-CHANGE and TARMAQ. The results find that random forest classification significantly outperforms previous approaches, receives lower Brier scores, and has superior trade-off between precision and recall. The results are consistent across software system and mining algorithm.</p>
Evaluating Software Modelling Recommendations: Towards Systematic Guidelines for Modelling
<p>The dataset contains four supporting documents for the paper title, "Evaluating Software Modelling Recommendations: Towards Systematic Guidelines for Modelling". </p> <p>The first documment is a pdf file containing the student survey questions.</p> <p>The second docummnet is an excel sheet containing the discussion of a focus group study. </p> <p>The third docummet is an excel sheet containing interview transcripts of 4 of the participants of this study (who agreed to publish their statements). We had to modify the transcript and omit a few discussions as it involves iedntification of names. <br><br><br>The final documment is a grading rubric used in one of the courses for grading. </p>
UU Webinar 3: Janja Komljenovic on Edtech policy recommendations and principles
<p>Title: Higher Education Edtech Policy Recommendations and Principles 1.0</p> <p>Abstract: This webinar will present the first version of policy recommendations for policymakers, universities, Edtech companies, and investors in Edtech. The recommendations are coming from the ESRC-funded research project on new forms of value in Edtech. </p> <p>Speaker: Janja Komljenovic</p> <p>Bio: Janja Komljenovic is a Senior Lecturer at Lancaster University in the UK. Her research focuses on the political economy of higher education. She is particularly interested in digital markets in higher education, new forms of value in digital education and the EdTech industry, and assetization as a new form of governance. Her approach sits at the intersection of economic sociology, science and technology studies, and higher education research. Janja is published internationally on higher education policy, markets, and educational technology.</p> <p>Link to Janja’s website: https://www.lancaster.ac.uk/education... </p> <p>Date of event: 22 June 2023</p>
Dataset Online grocery shopping recommender systems: common approaches and practices
<p>Standardized Excel form for data extraction. Extraction criteria were defined based on the sub-research questions that are provided in the paper. </p>
Data for "Uncertainty quantification in geochemical mapping: a review and recommendations"
<p>Data for "Uncertainty quantification in geochemical mapping: a review and recommendations".</p>
Figure 1 in Recreational watercraft decontamination: can current recommendations reduce aquatic invasive species spread?
Figure 1. Air-drying conditions within recommended durations, producing 100% mortality.
Data set in the form a relational database (sql) to denote a network of service providers, service clients and recommenders
<p>This data-set pertains to a network (i.e. graph) represented in the form of a relational data-base of service providers (nodes), service clients (nodes), service recommenders (nodes) and relationaships between then (i.e. a client used a provider, a recommender recommended a service to another client), along with some initial values of the QoS level perceived by any client whi have used a service and the reputation of a recommender. The data-set can be used for developing a reputation-based trust system. </p>
MESSAR: Automated recommendation of metabolite substructures from tandem mass spectra
<p>Data set associated with:</p> <p>Mrzic, A. et al. MESSAR: Automated recommendation of metabolite substructures from tandem mass spectra. bioRxiv (2017). doi:10.1101/134189</p> <p>This data set is provided by Janssen Pharmaceutica. It consists of known standard pharmaceutical compounds for which high quality Q-Exactive MS/MS data is provided. The data set contain drugs ranging from antifungal and antipsychotic agents to inhibitors of the hepatitis C virus and a compound that slows down progression of Alzheimer's disease.</p>
Kernelized rank learning for personalized drug recommendation (training and evaluation datasets)
<p>Training and evaluation datasets for our "Kernelized rank learning for personalized drug recommendation" paper:</p> <p>He* X, Folkman* L & Borgwardt K (2018), <em>Bioinformatics</em> <strong>34</strong>(16), 2808–2816, <a href="https://doi.org/10.1093/bioinformatics/bty132">https://doi.org/10.1093/bioinformatics/bty132</a><br> *equal contributions</p> <p>For the source code, please visit the GitHub repository: <a href="https://github.com/BorgwardtLab/Kernelized-Rank-Learning">https://github.com/BorgwardtLab/Kernelized-Rank-Learning</a></p> <p> </p>
PODCAST: The Impact of News Recommender Systems on our Personal Identity
<p>This 45-minute podcast is a result of qualitative research which aims to identify if we should be concerned that news recommender systems may have an impact on our personal identities over time. Ind sets out to achieve a number of objectives.</p> <p>1. To understand how news recommender systems influence the way of seeing and experiencing the world in which we anchor our identities<br> 2. To examine the extent to which news recommendations are transforming and shaping our preferences and behaviours<br> 3. To determine if we should be concerned about personal data privacy, as news recommendations technology collects our personal data to create algorithmically generated news recommendations<br> 4. To explore if we think that personalised news recommendations can serve to sharpen our focus and broaden our minds in parallel</p> <p>The Impact of News Recommender Systems on our Personal Identity is licensed under a Creative Commons License.</p>
Figure S1 in Development of experimental mesocosms for cicada nymphs Graptopsaltria nigrofuscata: methodology and research recommendations
Figure S1. Mean daily soil temperature during the mesocosm experiment (16 April to 6 July 2021).
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.