0.5.3

(#122) Add support for date/time features

0.5.2

Clip values in log-normal scale transform

0.5.1

disable shap for NN and CatBoost (#112, #114)

0.5.0

#100 Add modes: Explain, Perform, Compete
#108 Faster NN implementation for small datasets
#110 Fix warnings in permutation importance

0.4.1

Minor fixes to 0.4.0

0.4.0

Enhancements

• #96 Use internal early stopping
• #95 convergence warning in linear algorithm
• #94 Add support for kNN
• #90 Try to provide more meaningful names to models
• #74 Add stacked models
• #72 Add support for Neural Networks
• #70 Select best model after each iteration
• #64 Generate data information once
• #50 Add validation with split

Bugs

• #91 Dont run preprocessing for baseline algorithm
• #85 Cant load CatBoost model for predictions

0.3.5

Removed mae from sklearn decision tree based algorithms because they slow down the training

0.3.0

enhancement

• #83 Compare all models visually
• #79 Aggregate importances to one plot
• #78 Sort linear model coefficients
• #77 Shuffle generated models
• #75 Add SHAP explanations to models
• #69 Add skip_interpret argument
• #61 Add ExtraTrees
• #52 Add Linear and Logistic Regression support
• #51 Add LightGBM support
• #27 feature importance [enhancement]
• #68 Add minimum number of steps for algorithm

bug

• #66 Doesnt work with bytes as class
• #65 Dont generate parameters for Baseline
• #63 wrong time estimation for model training
• #62 Hill climbing not iterating on all algorithms
• #60 Show number of trees in learning curves

refactor

• #47 Refactor AutoML additional_metrics method

v0.1.6

#20 Add preprocessing for new data in case of missing values not present in train data.

v0.1.5

fix tqdm on jupyter