/Covers

Covers

Primary LanguageJupyter Notebook

Covers

Covers

unsupervised classification with kmeans of 96000 covers of books

  • classification file 'searching_k.csv' contains predictions for 3, 10, 33, 100, 333, 1000, 3333 clusters
  • classificatiof file 'searching_k2.csv' contains predictions for 100 clusters (column cat0) plus 100 subclusters for each of topclusters (column cat1). Predictions are aggregated in column cat-0-1 (100*cat0+cat1)