Python Automated Machine Learning tool :TPOT

TPOT是一个开源的机器学习项目，项目地址为：https://github.com/EpistasisLab/tpot

 

1. TPOT with code

<strong>step 1: 导入类模块</strong><br />from tpot import TPOTClassifier    #分类器
from tpot import TPOTRegressor     #回归器

<strong>step 2: 实例化（default)</strong><br />#创建默认分类器
default_pipeline_optimizer_classifier = TPOTClassifier()
#创建默认回归器
default_pipeline_optimizer_regressor  = TPOTRegressor()

<strong>step 2: 实例化（custom)</strong><br />#创建自定义分类器
custom_pipeline_optimezer_classifier = TPOTClassifier(generations=50,population_size=50,cv=5,random_state=100,
                                                      verbosity=2)
#创建自定义回归器
custom_pipeline_optimezer_regressor  =TPOTRegressor(generations=5,population_size=5,cv=5,random_state=20,
                                                      verbosity=1)

<strong>step 3</strong><strong>: 准备训练集、测试集</strong>
X_train, y_train, X_test, y_test = ?
#可以使用sklearn.model_selection.train_test_split()函数
<strong>
step 4</strong><strong>: 训练</strong>
custom_pipeline_optimezer_regressor.fit(X_train, y_train)

<strong>step </strong><strong>5: 测试
</strong>print(custom_pipeline_optimezer_regressor.score(X_test, y_test))

<strong>step </strong><strong>6: export the corresponding Python code for</strong><strong> the optimized pipeline</strong>
custom_pipeline_optimezer_regressor.export('tpot_exported_pipeline.py')

 2.scoring function

<strong>方式一：pass a string to the attribute <em>scoring</em></strong><br />属性值可以为<br /><em>'accuracy', 'adjusted_rand_score', 'average_precision', 'balanced_accuracy',</em><br /><em>'f1','f1_macro', 'f1_micro', 'f1_samples', 'f1_weighted', 'neg_log_loss', 'neg_mean_absolute_error',
'neg_mean_squared_error', 'neg_median_absolute_error', 'precision', 'precision_macro', 'precision_micro',
'precision_samples', 'precision_weighted','r2', 'recall', 'recall_macro', 'recall_micro', 'recall_samples',
'recall_weighted', 'roc_auc', 'my_module.scorer_name*'</em><br />

<strong>方式二：<em>用户自定义</em></strong><br /># Make a custom metric function
def my_scoring_func(y_true, y_pred):
    return mean_squared_error(y_true, y_pred)

# Make a custom a scorer from the custom metric function
# Note: greater_is_better=False in make_scorer below would mean that the scoring function should be minimized.
my_scorer = sklearn.metrics.scorer.make_scorer(my_scoring_func,greater_is_better=False)<br />custom_pipeline_optimezer_regressor  =TPOTRegressor(generations=5,population_size=5,cv=5,random_state=20,
                                                      verbosity=1,scoring=my_scorer)

3.config_dict

有四个默认的configuration options

1. Default TPOT
2. TPOT light
3. TPOT MDR
4. TPOT sparse

具体说明：http://epistasislab.github.io/tpot/using/#built-in-tpot-configurations

custom_pipeline_optimezer_regressor  =TPOTRegressor(generations=5,population_size=5,cv=5,random_state=20,
                                                      verbosity=1,config_dict='TPOT light')

 

4.用户自定义config

tpot_config = {
    'sklearn.naive_bayes.GaussianNB': {
    },

    'sklearn.naive_bayes.BernoulliNB': {
        'alpha': [1e-3, 1e-2, 1e-1, 1., 10., 100.],
        'fit_prior': [True, False]
    },

    'sklearn.naive_bayes.MultinomialNB': {
        'alpha': [1e-3, 1e-2, 1e-1, 1., 10., 100.],
        'fit_prior': [True, False]
    }
}

custom_pipeline_optimezer_regressor  =TPOTRegressor(generations=5,population_size=5,cv=5,random_state=20,
                                                      verbosity=1,config_dict=tpot_config)

 5.分布式环境训练

from sklearn.externals import joblib
import distributed.joblib
from dask.distributed import Client

# connect to the cluster
client = Client('schedueler-address')

# create the estimator normally
estimator = TPOTClassifier(n_jobs=-1)

# perform the fit in this context manager
with joblib.parallel_backend("dask"):
    estimator.fit(X, y)

 6.实际项目

项目目标是预测下游水库的进水量，其源数据内容如下

第一列是下游水库的进水量，第二列是上游水库的出水量，其余的是上下游之间降雨观测点的雨量信息 

【说明】：本文章由站长整理发布，文章内容不代表本站观点，如文中有侵权行为，请与本站客服联系(QQ:254677821)！