https://github.com/nicomignoni/tiresia

Tiny sklearn-based AutoML tool.
https://github.com/nicomignoni/tiresia

automl data-science machine-learning sklearn

Last synced: about 2 months ago
JSON representation

Tiny sklearn-based AutoML tool.

• Host: GitHub
• URL: https://github.com/nicomignoni/tiresia
• Owner: nicomignoni
• License: mit
• Created: 2021-01-22T16:52:03.000Z (about 4 years ago)
• Default Branch: master
• Last Pushed: 2021-02-11T17:00:35.000Z (almost 4 years ago)
• Last Synced: 2024-12-15T22:38:15.399Z (about 2 months ago)
• Topics: automl, data-science, machine-learning, sklearn
• Language: Python
• Homepage:
• Size: 303 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        


# tiresia

[![PyPI version shields.io](https://img.shields.io/pypi/v/tiresia.svg)](https://pypi.python.org/pypi/tiresia/)

## Installation

```

pip install tiresia

```

## Description

[tiresia](https://en.wikipedia.org/wiki/Tiresias) is just a wrapper around [scikit-learn](https://scikit-learn.org/stable/) ```GridSearchCV```. The idea is to simplify the model testing workflow. With tiresia you can choose which models to test and to exclude and provide a ```param_grid``` for the ones you want to explore deeper, while keeping the parameters of the less interesting ones on default. 

## Example

```python

from tiresia.predictor import AutoPredictor

from sklearn.datasets import make_regression, make_classification

from sklearn.model_selection import train_test_split

from sklearn.metrics import roc_auc_score, r2_score

test_type = "classifier"

if test_type == "classifier":

    train, target = make_classification(1000, 20)

elif test_type == "regressor":

    train, target = make_regression(1000, 20)

    

x_train, x_test, y_train, y_test = train_test_split(train, target, test_size=0.3)

autopred = AutoPredictor(estimator_type=test_type)

        

autopred.fit(x_train, y_train, x_test, y_test, scoring=roc_auc_score, greater_is_better=True)

predictions = autopred.predictions

results     = autopred.results

print(results)

                       CLASSIFIER roc_auc_score

4          DecisionTreeClassifier         0.990

10     GradientBoostingClassifier         0.990

22         RandomForestClassifier         0.973

0              AdaBoostClassifier         0.963

1               BaggingClassifier         0.960

7            ExtraTreesClassifier         0.960

24              RidgeClassifierCV         0.936

23                RidgeClassifier         0.936

18                          NuSVC         0.936

14     LinearDiscriminantAnalysis         0.936

16             LogisticRegression         0.923

26                            SVC         0.923

3          CalibratedClassifierCV         0.920

15                      LinearSVC         0.917

17                NearestCentroid         0.917

25                  SGDClassifier         0.916

6             ExtraTreeClassifier         0.907

19    PassiveAggressiveClassifier         0.907

20                     Perceptron         0.906

11           KNeighborsClassifier         0.893

8                      GaussianNB         0.884

9       GaussianProcessClassifier         0.870

2                     BernoulliNB         0.840

12               LabelPropagation         0.827

13                 LabelSpreading         0.827

21  QuadraticDiscriminantAnalysis         0.615

5                 DummyClassifier         0.473

```