Click to add a brief description of the dataset (Markdown and LaTeX enabled).

Provide:

a high-level explanation of the dataset characteristics explain motivations and summary of its content potential use cases of the dataset

The AMiner Dataset is a collection of different relational datasets. It consists of a set of relational networks such as citation networks, academic social networks or topic-paper-autor networks among others.

This dataset is a knowledge graph extracted from a AMiner benchmark for a research project on knowledge graph embeddings (KGEs) for author disambiguation. Structural triples of the knowledge graph are split into training, testing and validation for applying representation learning methods. Textual literals and numeric literals were stored separately in order to implement multimodal approaches for KGEs (see arXiv:1802.00934). For the same reason, textual literals and numeric literals are already stored into sentence embeddings and a numeric matrix respectively in the files textual_literals.npy and numeric_literals.npy. The file and_eval.json contains the evaluation dataset used for evaluating our AND architecture. For the script used to gather this dataset see the GitHub repository: https://github.com/sntcristian/and-kge/tree/main/aminer.

This dataset is a knowledge graph extracted from a AMiner benchmark for a research project on knowledge graph embeddings (KGEs) for author disambiguation. Structural triples of the knowledge graph are split into training, testing and validation for applying representation learning methods. Textual literals and numeric literals were stored separately in order to implement multimodal approaches for KGEs (see arXiv:1802.00934). For the same reason, textual literals and numeric literals are already stored into sentence embeddings and a numeric matrix respectively in the files textual_literals.npy and numeric_literals.npy. For the script used to gather this dataset see the GitHub repository: https://github.com/sntcristian/and-kge/tree/main/aminer.

AMiner (aminer.org) aims to provide comprehensive search and mining services for researcher social networks. The system focuses on: (1) creating a semantic-based profile for each researcher by extracting information from the distributed Web; (2) integrating academic data (e.g., the bibliographic data and the researcher profiles) from multiple sources; (3) accurately searching the heterogeneous network; (4) analyzing and discovering interesting patterns from the built researcher social network. The main search and analysis functions in AMiner include: profile search, expert finding, conference analysis, course search, sub-graph search, topic browser, academic ranks, and user management.

This data is derived from Aminer's publicly available disambiguation dataset (https://open.aminer.cn/article?id=55af4228dabfae1ce3ed1253), on the basis of which it explores how the characterization of research collaborations affects the novelty and impact of knowledge outcomes

8 figures of the paper. Figure 1 presents the architecture of AMiner. Figure 2 shows the schema of the researcher profile. Figure 3 is an example of researcher profile. Figure 4 is an overview of the name disambiguation framework in AMiner. Figure 5 is graphical representation of the three Author-Conference-Topic (ACT) models. Figure 6 shows an example result of experts found for “Data Mining”. Figure 7 is a model framework of DeepInf. Figure 8 shows an example of researcher ranking by sociability index.

This dataset consists in two distinct scholarly knowledge graph created from two publicly available bibliographic datasets: 1) a triplestore covering information about the journal Scientometrics provided by OpenCitations (available here), and 2) the AMiner AND benchmark from 2018 available here. This KG was extracted for a research project on knowledge graph embeddings (KGEs) for author disambiguation. Structural triples of the knowledge graphs are split into training, testing and validation for applying representation learning methods. Textual literals and numeric literals were stored separately in order to implement multimodal approaches for KGEs (see arXiv:1802.00934). For the same reason, textual literals and numeric literals are already stored into sentence embeddings and a numeric matrix respectively in the files textual_literals.npy and numeric_literals.npy in order to simplify the representation learning task. The file and_eval.json of each KG contains the evaluation dataset used for evaluating our AND architecture. For the script used to gather this dataset see https://github.com/sntcristian/and-kge/tree/main/src/AMiner-534K and https://github.com/sntcristian/and-kge/tree/main/src/OC-782K.

This dataset is a knowledge graph extracted from a triplestore covering information about the journal Scientometrics and modelled according to the OpenCitations Data Model. The original triplestore is available here. This KG was extracted for a research project on knowledge graph embeddings (KGEs) for author disambiguation. Structural triples of the knowledge graph are split into training, testing and validation for applying representation learning methods. Textual literals and numeric literals were stored separately in order to implement multimodal approaches for KGEs (see arXiv:1802.00934). For the same reason, textual literals and numeric literals are already stored into sentence embeddings and a numeric matrix respectively in the files textual_literals.npy and numeric_literals.npy. The file and_eval.json contains the evaluation dataset used for evaluating our AND architecture. For the script used to gather this dataset see the GitHub repository: https://github.com/sntcristian/and-kge/tree/main/open-citations.

This dataset contains information about academic articles, their authors and venues of publication. The dataset has the form of a graph. It has been produced by the SmartDataLake project (https://smartdatalake.eu), using data collected from Aminer (https://aminer.org).

A copy of the "Open Academic Graph v2" (OAGv2) corpus published by aminer.org and Microsoft Academic Graph in early 2019. Contains roughly 90 GB (compressed) of bibliographic metadata for hundreds of millions of publications. Related publications include: Jie Tang, Jing Zhang, Limin Yao, Juanzi Li, Li Zhang, and Zhong Su. ArnetMiner: Extraction and Mining of Academic Social Networks. In Proceedings of the Fourteenth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (SIGKDD 2008). pp.990-998. Arnab Sinha, Zhihong Shen, Yang Song, Hao Ma, Darrin Eide, Bo-June (Paul) Hsu, and Kuansan Wang. 2015. An Overview of Microsoft Academic Service (MAS) and Applications. In Proceedings of the 24th International Conference on World Wide Web (WWW ’15 Companion). ACM, New York, NY, USA, 243-246.

OAGT is a paper topic dataset consisting of 6942930 records which comprise various scientific publication attributes like abstracts, titles, keywords, publication years, venues, etc. The last two fields of each record are the topic id from a taxonomy of 27 topics created from the entire collection and the 20 most significant topic words. Each dataset record (sample) is stored as a JSON line in the text file.

The data is derived from OAG data collection (https://aminer.org/open-academic-graph) which was released
under ODC-BY license.

This data (OAGT Paper Topic Dataset) is released under CC-BY license (https://creativecommons.org/licenses/by/4.0/).

If using it, please cite the following paper:

Erion Çano, Benjamin Roth: Topic Segmentation of Research Article Collections. ArXiv 2022, CoRR abs/2205.11249, https://doi.org/10.48550/arXiv.2205.11249

Comparison of this paper’s method with ten baseline methods on the Aminer dataset for three types of metrics.

Taken from here https://www.aminer.org/citation and converted to csv (but why)

DBLP (https://dblp.org/) is a comprehensive collection of computer science publications from major and minor journals and conference proceedings. From this dump, we remove arXiv preprints. Our dataset consists of 1.9 million publications from 1970 to 2014 that are authored by 1.1 million authors. We have added citations among publications by combining DBLP with the AMiner dataset (https://www.aminer.org/citation) via publication titles and years. There are 6.6 million citations among publications. Author names in DBLP are disambiguated. To infer the gender of authors, we have used a method that combines the results of name-based and image-based gender detection services. Since the accuracy is very low for Chinese and Korean names, we label their gender as unknown to reduce noise in our analysis.

OAGL is a paper length prediction dataset consisting of 17528680 records which comprise various scientific publication metadata like abstracts, titles, keywords, publication years, venues, etc. The last field of each record is the page length of the corresponding publication. Dataset records (samples) are stored as JSON lines in each text file. The data is derived from OAG data collection (https://aminer.org/open-academic-graph) which was released under ODC-BY license. This data (OAGL Paper Length Dataset) is released under CC-BY license (https://creativecommons.org/licenses/by/4.0/).

With the dramatic increase in the number of published papers and the continuous progress of deep learning technology, the research on name disambiguation is at a historic peak, the number of paper authors is increasing every year, and the situation of authors with the same name is intensifying, therefore, it is a great challenge to accurately assign the newly published papers to their respective authors. The current mainstream methods for author disambiguation are mainly divided into two methods: feature-based clustering and connection-based clustering, but none of the current mainstream methods can efficiently deal with the author name disambiguation problem, For this reason, this paper proposes the author name ablation method based on the relational graph heterogeneous attention neural network, first extract the semantic and relational information of the paper, use the constructed graph convolutional embedding module to train the splicing to get a better feature representation, and input the constructed network to get the vector representation. As the existing graph heterogeneous neural network can not learn different types of nodes and edge interaction, add multiple attention, design ablation experiments to verify its impact on the network. Finally improve the traditional hierarchical clustering method, combined with the graph relationship and topology, using training vectors instead of distance calculation, can automatically determine the optimal k-value, improve the accuracy and efficiency of clustering. The experimental results show that the average F1 value of this paper’s method on the Aminer dataset is 0.834, which is higher than other mainstream methods.

PERSON Dataset V2:Dataset created for paper "Search Personalization Based on Social-Network-Based Interestedness Measures." Please cite the paper for any usage.The dataset is produced by data cleaning of AMiner's citation network V2 dataset (https://aminer.org/citation). Anyone who wants to use PERSON V2 dataset must cite Aminer's dataset (as explained in its homepage: Jie Tang, Jing Zhang, Limin Yao, Juanzi Li, Li Zhang, and Zhong Su. ArnetMiner: Extraction and Mining of Academic Social Networks. In Proceedings of the Fourteenth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (SIGKDD'2008). pp.990-998) as well as the aforementioned paper.It includes two files: 1- authors_giant.txt: the information of authors and their co-authors. The format is as follows: author ID author name
the list of coauthors delimited by "," (Each entry contains the ID of the coauthor followed by the number of times they co-authored a paper) ... 2- papers_giant.txt: the information of papers and references. The format is as follows: paper ID Is paper merged (See the first paper for details) original paper ID (in Aminer's dataset) blank blank blank blank title abstract time (only the year part is important) blank references to papers out of the PERSON dataset (indicated by Aminer's IDs) references to papers inside the PERSON dataset (indicated by PERSON's IDs) author IDs ...

OAGL is a paper metadata dataset consisting of 17528680 records which comprise various scientific publication attributes like abstracts, titles, keywords, publication years, venues, etc. The last field of each record is the page length of the corresponding publication. Dataset records (samples) are stored as JSON lines in each text file. The data is derived from OAG data collection (https://aminer.org/open-academic-graph) which was released under ODC-BY license. This data (OAGL Paper Metadata Dataset) is released under CC-BY license (https://creativecommons.org/licenses/by/4.0/). If using it, please cite the following paper:

Çano Erion, Bojar Ondřej: How Many Pages? Paper Length Prediction from the Metadata. NLPIR 2020, Proceedings of the the 4th International Conference on Natural Language Processing and Information Retrieval, Seoul, Korea, December 2020.

This data set contains citation network data for 1320 publications from dblp (https://dblp.uni-trier.de/) enriched with data from AMiner (https://aminer.org/) for classification of seminal and survey publications.

Citations and references are contained for every publication. For each of the 121,084 papers, dblp key, publication year as well as stemmed and unstemmed concatenations of its title and abstract are given. Seminal papers come from A* conferences, surveys were extracted from venues specialized in publishing reviews.

For details, see Revaluating Semantometrics from Computer Science Publications, Christin Katharina Kreutz, Premtim Sahitaj, and Ralf Schenkel, 2019, submitted to BIRNDL@SIGIR.