Xstructure is a service for browsing and searching papers in arxiv.org

Among the features of this service are:

  • Automated generation of hierarchical classification scheme for the papers. The scheme results from classification of the papers in the arxiv database. The only input for the classification is the citation graph. The number of the levels in the hierarchy and the number of the clusters is determined by the algorithm. The algorithm creates the classification scheme, and indexes the papers by the created classification;
  • The classification is used to index the new papers. We plan to rebuild the classification scheme regularly. In this way, we will take into account that appearance of new papers may lead to emergence of new themes. Detection of new themes is one of our objectives;
  • A number of extra attributes (e.g. Theme name, Authority and Review Articles, etc.) for the elements (themes) of the classification  (see Help);
  • Accessability of the classification in response to search requests via display options, e.g., display as Tree of Themes, and Refrerence (Citation) Tree.

About 10% of papers from arxiv are missed in our database. We work on decreasing this number.

Comments, questions, and suggestions are to be sent to Grigorii Pivovarov