You are currently on a failover version of the Materials Cloud Archive hosted at CINECA, Italy.
Click here to access the main Materials Cloud Archive.
Note: If the link above redirects you to this page, it means that the Archive is currently offline due to maintenance. We will be back online as soon as possible.
This version is read-only: you can view published records and download files, but you cannot create new records or make changes to existing ones.

Machine learning on multiple topological materials datasets


JSON Export

{
  "metadata": {
    "edited_by": 1641, 
    "references": [
      {
        "type": "Website", 
        "citation": "Yuqing He et al., to be submitted", 
        "url": "topoclass.modl-uclouvain.org"
      }, 
      {
        "comment": "Paper describing the work performed", 
        "type": "Journal reference", 
        "citation": "Yuqing He at al., in preparation"
      }
    ], 
    "license_addendum": null, 
    "status": "published", 
    "is_last": false, 
    "contributors": [
      {
        "affiliations": [
          "UCLouvain, Institut de la Matiere Condens\u00e9e et des Nanosciences (IMCN), Chemin des \u00c9toiles 8, Louvain-la-Neuve 1348, Belgium", 
          "Beijing National Laboratory for Condensed Matter Physics and Institute of Physics,Chinese Academy of Sciences, Beijing, China"
        ], 
        "givennames": "Yuqing", 
        "familyname": "He"
      }, 
      {
        "affiliations": [
          "UCLouvain, Institut de la Matiere Condens\u00e9e et des Nanosciences (IMCN), Chemin des \u00c9toiles 8, Louvain-la-Neuve 1348, Belgium"
        ], 
        "givennames": "Pierre-Paul", 
        "familyname": "De Breuck"
      }, 
      {
        "affiliations": [
          "Beijing National Laboratory for Condensed Matter Physics and Institute of Physics,Chinese Academy of Sciences, Beijing, China"
        ], 
        "givennames": "Hongming", 
        "familyname": "Weng"
      }, 
      {
        "affiliations": [
          "UCLouvain, Institut de la Matiere Condens\u00e9e et des Nanosciences (IMCN), Chemin des \u00c9toiles 8, Louvain-la-Neuve 1348, Belgium"
        ], 
        "givennames": "Matteo", 
        "familyname": "Giantomassi"
      }, 
      {
        "email": "gian-marco.rignanese@uclouvain.be", 
        "givennames": "Gian-Marco", 
        "familyname": "Rignanese", 
        "affiliations": [
          "UCLouvain, Institut de la Matiere Condens\u00e9e et des Nanosciences (IMCN), Chemin des \u00c9toiles 8, Louvain-la-Neuve 1348, Belgium"
        ]
      }
    ], 
    "doi": "10.24435/materialscloud:xx-xb", 
    "conceptrecid": "2555", 
    "owner": 1641, 
    "_oai": {
      "id": "oai:materialscloud.org:2556"
    }, 
    "version": 1, 
    "keywords": [
      "topological materials", 
      "machine learning", 
      "database"
    ], 
    "description": "A dataset of 35,608 materials with their topological properties is constructed by combining the density functional theory (DFT) results of Materiae and the Topological Materials Database. Thanks to this, machine-learning approaches are developed to categorize materials into five distinct topological types, with the XGBoost model achieving an impressive 85.2% classification accuracy. By conducting generalization tests on different sub-datasets, differences are identified between the original datasets in terms of topological types, chemical elements, unknown magnetic compounds, and feature space coverage. Their impact on model performance is analyzed.\nTurning to the simpler binary classification between trivial insulators and nontrivial topological materials, three different approaches are also tested. Key characteristics influencing material topology are identified, with the maximum packing efficiency and the fraction of p valence electrons being highlighted as critical features", 
    "_files": [
      {
        "size": 49935, 
        "checksum": "md5:e97367839d6f923849f6ab81ff7830f8", 
        "description": "Python notebook to analyze the data", 
        "key": "topoclass.ipynb"
      }, 
      {
        "size": 540662618, 
        "checksum": "md5:57066e24446f9f128c63a8e1698880de", 
        "description": "JSON gzipped file with all the data", 
        "key": "topoclass.json.gz"
      }, 
      {
        "size": 20722891, 
        "checksum": "md5:d9c6901e2781b55382233974c5360bfc", 
        "description": "Chemiscope visualization of the dataset, that can be used interactively on Materials Cloud", 
        "key": "topoclass.chemiscope.json.gz"
      }, 
      {
        "size": 106726120, 
        "checksum": "md5:2859bad2900c0ad1057e8b3d357f452e", 
        "description": "OPTIMADE file with all the data", 
        "key": "optimade.jsonl"
      }
    ], 
    "publication_date": "Feb 12, 2025, 09:58:11", 
    "license": "Creative Commons Attribution 4.0 International", 
    "title": "Machine learning on multiple topological materials datasets", 
    "mcid": "2025.27", 
    "id": "2556"
  }, 
  "revision": 8, 
  "created": "2025-02-06T18:06:52.738034+00:00", 
  "id": "2556", 
  "updated": "2025-02-26T13:07:05.560471+00:00"
}