Task: Building a Knowledge Graph and Recommender System for Research Papers
Budget: ₹600 – ₹1,500 INR
Problem Statement
Knowledge graphs have emerged as powerful tools for representing and organizing
information in a structured and interconnected manner. They enable a deeper
understanding of complex domains by capturing relationships and contextual
information between entities. By representing knowledge as a graph, with entities as
nodes and relationships as edges, knowledge graphs provide a holistic view of the data,
facilitating efficient data integration, knowledge discovery, and advanced analytics.
Dataset link: https://bit.ly/balkanid-ds-dataset-download
Part 1: Charting the Research Landscape
You are part of a team of researchers who have embarked on a quest to uncover
hidden knowledge and insights within a vast collection of research papers. As
you journey through the realm of research, you discover the importance of
organizing and modeling metadata effectively. Utilizing the RDF framework and
the expressive SPARQL language, build a knowledge graph that represents the
interconnectedness of research papers, their authors, citations, and other
relevant metadata. Showcase your skills by creating a powerful visualization tool
that enables researchers to visualize the results of SPARQL queries as a
subgraph.
Note: Draw a diagram to illustrate the ontology of your knowledge graph
depicting entities and their relationships using a diagramming software (like
lucidcharts, powerpoint, draw.io, etc.). Also, provide at least 5 examples of the
visualizations generated by executing SPARQL queries using the visualization
tool that you have built.
Part 2: The Search for Knowledge
To aid fellow researchers in their exploration, your task is to design and
implement a recommender system that suggests 5 similar and important
research papers that a researcher could cite for a new research paper that
he/she plans to write. Embrace the power of graph-related algorithms and
learning models to generate meaningful recommendations. If implementing graph
algorithms for recommendations proves challenging, use any other algorithm, but
provide a detailed description of the steps you would take to incorporate graph
algorithms in the recommender system.
In new_research_papers.jsonl (https://bit.ly/balkanid-new-data-download) we
have provided the title, discipline and abstract of 3 papers that the researcher
plans to write. Using your recommender system, for each new research paper,
recommend 5 similar and important papers that the researcher could cite (in
order).
Note: Importance of a research paper is measured not just by the number of
citations that it receives but also by the quality of the citations. The quality of a
citation is in turn determined by the number and quality of its own citations.
Knowledge graphs have emerged as powerful tools for representing and organizing
information in a structured and interconnected manner. They enable a deeper
understanding of complex domains by capturing relationships and contextual
information between entities. By representing knowledge as a graph, with entities as
nodes and relationships as edges, knowledge graphs provide a holistic view of the data,
facilitating efficient data integration, knowledge discovery, and advanced analytics.
Dataset link: https://bit.ly/balkanid-ds-dataset-download
Part 1: Charting the Research Landscape
You are part of a team of researchers who have embarked on a quest to uncover
hidden knowledge and insights within a vast collection of research papers. As
you journey through the realm of research, you discover the importance of
organizing and modeling metadata effectively. Utilizing the RDF framework and
the expressive SPARQL language, build a knowledge graph that represents the
interconnectedness of research papers, their authors, citations, and other
relevant metadata. Showcase your skills by creating a powerful visualization tool
that enables researchers to visualize the results of SPARQL queries as a
subgraph.
Note: Draw a diagram to illustrate the ontology of your knowledge graph
depicting entities and their relationships using a diagramming software (like
lucidcharts, powerpoint, draw.io, etc.). Also, provide at least 5 examples of the
visualizations generated by executing SPARQL queries using the visualization
tool that you have built.
Part 2: The Search for Knowledge
To aid fellow researchers in their exploration, your task is to design and
implement a recommender system that suggests 5 similar and important
research papers that a researcher could cite for a new research paper that
he/she plans to write. Embrace the power of graph-related algorithms and
learning models to generate meaningful recommendations. If implementing graph
algorithms for recommendations proves challenging, use any other algorithm, but
provide a detailed description of the steps you would take to incorporate graph
algorithms in the recommender system.
In new_research_papers.jsonl (https://bit.ly/balkanid-new-data-download) we
have provided the title, discipline and abstract of 3 papers that the researcher
plans to write. Using your recommender system, for each new research paper,
recommend 5 similar and important papers that the researcher could cite (in
order).
Note: Importance of a research paper is measured not just by the number of
citations that it receives but also by the quality of the citations. The quality of a
citation is in turn determined by the number and quality of its own citations.