v1.1.0

• Update models #88
  • Create SVM
  • Create KNN
  • Create RandomForest
  • Update NN (TF keras)
• Show feature importance for all possible models #89
  • CatBoost
  • XGBoost
  • Ridge
  • Random Forest

v1.0.7

• Fix ResRunner
• Fix starter.sh
• Fix XGBoost verbose
• Update README.md

v1.0.6

• Create model_nn #79
• Change input path to create pickle folder #80

v1.0.5

• Update docker image from v76 to v88 close #77
• Docker mlflow settings #76
• Solve conflicts between black and isort

v1.0.4

Fix FeatureStore

v1.0.3

Introduce flake8, isort, black, and mypy.

v1.0.2

Fix model.runner.

v1.0.1

Fix target encoding.

v1.0.0

This is the release note of v1.0.0.

Highlights

The concept of the first major version is "Keep it simple".

preprocessing

• Delete too complicated original pipelines for feature engineerings
• Instead, use xfeat for simplification
• Delete text processing because now we can refer nlp-recipes-ja if necessary

model

• Delete unfamiliar algorithm such as NGBoost, TabNet, and TNN

v0.1.0

This is the release note of v0.1.0.