v0.9.15

What's Changed

• Fix AutoLGB. Reformat with black by @jeongyoonlee in #71
• up the version to 0.9.15 by @jeongyoonlee in #73
• Update test.yml by @jeongyoonlee in #69
• Update python-publish.yml by @jeongyoonlee in #70

Full Changelog: v0.9.14...v0.9.15

v0.9.14

What's Changed

• Update python-publish.yml by @jeongyoonlee in #64
• add plot_curve() for plotting ROC and PR curves by @jeongyoonlee in #66
• fix build error by replacing ml_metrics's kappa with scikit-learn's by @jeongyoonlee in #67
• up the version to 0.9.14 by @jeongyoonlee in #68

Full Changelog: v0.9.13...v0.9.14

v0.9.13

• add transfer learning with the pretrained_model input argument 
• allow to set the learning_rate in __init__()
• add a test for transfer learning between DAE/SDAE

v0.9.12

• make label_encoding=True default in DAE/SDAE

v0.9.11

• fix an error raised when printing out the DAE/SDAE objects
• update random_state/seed arguments in DAE/SDAE/DAELayer to follow scikit-learn/tensorflow conventions
• up the version to v0.9.11

v0.9.10

• add options to add more than 1 encoder in DAELayer
• add options to add validation_data in DAE/SDAE
• make label-encoding optional in DAE/SDAE

v0.9.9

• fix #56: DAE/SDAE's transform() changes the input dataframe.

v0.9.8

• hotfix SDAE

v0.9.7

• add SDAE, supervised denoising AutoEncoder

v0.9.5

• copy dataframe before transforming it in encoders to prevent overwriting
• update the default threshold for feature selection in automl
• fix DAE with all numeric features