IT ticket category classifier

This artefact predicts one of eight IT help desk categories from free-text ticket content. It contains a TF-IDF and LinearSVC pipeline with calibrated confidence scores; transparent priority rules are separate from this artefact.

Results

  • Test accuracy: 0.8556
  • Test macro F1: 0.8575
  • Temperature-scaled log loss: 0.8951
  • Temperature-scaled calibration error: 0.3716
  • Confidence >= 0.8 coverage: 74.1%
  • Accuracy at confidence >= 0.8: 94.5%
  • Calibration temperature: 0.277762

Library versions

  • python: 3.13.16
  • pandas: 2.3.3
  • numpy: 2.1.3
  • scikit_learn: 1.6.1
  • xgboost: 3.4.1

Limitations

  • The data comes from one help desk and may not represent other organisations.
  • Priority is a separate rule layer and is not part of this model artefact.
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support