DoubleML

https://docs.doubleml.org/
Admin
DoubleML is an open-source Python and R library providing an implementation of the double/debiased...
Tokens:1,285,866
Snippets:5,229
Trust Score:6.4
Update:1 day ago
Show doc for...
Tokens:
Raw
### Import libraries and setup for TabPFN with DoubleML

Source: https://docs.doubleml.org/dev/examples/learners/py_tabpfn.html

Imports required libraries for the example, including TabPFN, DoubleML, scikit-learn, LightGBM, and data generation utilities. Suppresses specific warnings related to CPU usage and feature names. Requires separate TabPFN installation; CUDA-enabled PyTorch recommended for GPU acceleration.

```python
import numpy as np
import pandas as pd

import matplotlib.pyplot as plt
import seaborn as sns
import time

from sklearn.ensemble import RandomForestRegressor, RandomForestClassifier
from sklearn.linear_model import LinearRegression, LogisticRegression
import lightgbm as lgbm
from tabpfn import TabPFNRegressor, TabPFNClassifier
from tabpfn.constants import ModelVersion

import doubleml as dml
from doubleml.irm.datasets import make_irm_data_discrete_treatments

import warnings
warnings.filterwarnings("ignore", message="Running on CPU*", category=UserWarning, module="tabpfn")
warnings.filterwarnings("ignore", message=".*does not have valid feature names.*", category=UserWarning)
```

--------------------------------

### DoubleMLPLR with Optuna Tuning Example

Source: https://docs.doubleml.org/dev/api/generated/doubleml.did.DoubleMLDID.html

Complete example showing how to use DoubleMLPLR with LightGBM regressors and Optuna for hyperparameter tuning. It includes data generation, model setup, parameter grid functions, tuning with TPE sampler, and fitting the model to get summary results.

```python
>>> import numpy as np
>>> from doubleml import DoubleMLData, DoubleMLPLR
>>> from doubleml.plm.datasets import make_plr_CCDDHNR2018
>>> from lightgbm import LGBMRegressor
>>> import optuna
>>> # Generate data
>>> np.random.seed(42)
>>> data = make_plr_CCDDHNR2018(n_obs=500, dim_x=20, return_type='DataFrame')
>>> dml_data = DoubleMLData(data, 'y', 'd')
>>> # Initialize model
>>> dml_plr = DoubleMLPLR(
...    dml_data,
...    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42),
...    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42)
... )
>>> # Define parameter grid functions
>>> def ml_l_params(trial):
...     return {
...         'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
...     }
>>> def ml_m_params(trial):
...     return {
...         'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
...     }
>>> ml_param_space = {'ml_l': ml_l_params, 'ml_m': ml_m_params}
>>> # Tune with TPE sampler
>>> optuna_settings = {
...     'n_trials': 5,
...     'sampler': optuna.samplers.TPESampler(seed=42),
... }
>>> tune_res = dml_plr.tune_ml_models(ml_param_space, optuna_settings=optuna_settings, return_tune_res=True)
>>> print(tune_res[0]['ml_l'].best_params)  
{'learning_rate': 0.03907122389107094}
>>> # Fit and get results
>>> dml_plr.fit().summary 
      coef   std err          t         P>|t|     2.5 %    97.5 %
d  0.57436  0.045206  12.705519  5.510257e-37  0.485759  0.662961
>>> # Example with scoring methods and directions
>>> scoring_methods = {
...     'ml_l': 'neg_mean_squared_error',  # Negative metric
...     'ml_m': 'neg_mean_squared_error'
... }
>>> optuna_settings = {
...     'n_trials': 50,
...     'direction': 'maximize',  # Maximize negative MSE (minimize MSE)
...     'sampler': optuna.samplers.TPESampler(seed=42),
... }
>>> tune_res = dml_plr.tune_ml_models(ml_param_space, scoring_methods=scoring_methods,
...                                   optuna_settings=optuna_settings, return_tune_res=True)
>>> print(tune_res[0]['ml_l'].best_params)  
{'learning_rate': 0.04300012336462904}
>>> dml_plr.fit().summary 
       coef   std err          t         P>|t|     2.5 %    97.5 %
d  0.574796  0.045062  12.755721  2.896820e-37  0.486476  0.663115
```

--------------------------------

### Setup DoubleML data for tuning example

Source: https://docs.doubleml.org/dev/guide/learners.html

Initializes a DoubleMLData object from simulated data with 200 observations and 200 variables. Required before tuning a DoubleMLPLR model.

```R
library(DoubleML)
library(mlr3)
library(data.table)

set.seed(3141)
n_obs = 200
n_vars = 200
theta = 3
X = matrix(stats::rnorm(n_obs * n_vars), nrow = n_obs, ncol = n_vars)
d = X[, 1:3, drop = FALSE] %*% c(5, 5, 5) + stats::rnorm(n_obs)
y = theta * d + X[, 1:3, drop = FALSE] %*% c(5, 5, 5)  + stats::rnorm(n_obs)
dml_data = double_ml_data_from_matrix(X = X, y = y, d = d)
```

--------------------------------

### DoubleMLPQ Constructor and Example

Source: https://docs.doubleml.org/dev/api/generated/doubleml.irm.DoubleMLPQ.html

Example of creating a DoubleMLPQ object, fitting it, and displaying the summary of results.

```APIDOC
## DoubleMLPQ

### Description
Class for partially linear quantile regression models based on double machine learning.

### Usage
```python
import numpy as np
import doubleml as dml
from doubleml.irm.datasets import make_irm_data
from sklearn.ensemble import RandomForestClassifier

np.random.seed(3141)
ml_g = RandomForestClassifier(n_estimators=100, max_features=20, max_depth=10, min_samples_leaf=2)
ml_m = RandomForestClassifier(n_estimators=100, max_features=20, max_depth=10, min_samples_leaf=2)
data = make_irm_data(theta=0.5, n_obs=500, dim_x=20, return_type='DataFrame')
obj_dml_data = dml.DoubleMLData(data, 'y', 'd')
dml_pq_obj = dml.DoubleMLPQ(obj_dml_data, ml_g, ml_m, treatment=1, quantile=0.5)
dml_pq_obj.fit().summary
```

### Output Example
```
       coef   std err         t     P>|t|     2.5 %    97.5 %
d  0.553878  0.149858  3.696011  0.000219  0.260161  0.847595
```

### Methods
- `bootstrap(method, n_rep_boot)` - Multiplier bootstrap for DoubleML models.
- `confint(joint, level)` - Confidence intervals for DoubleML models.
- `construct_framework()` - Construct a `doubleml.DoubleMLFramework` object.
- `draw_sample_splitting()` - Draw sample splitting for DoubleML models.
- `evaluate_learners(learners, metric)` - Evaluate fitted learners for DoubleML models on cross-validated predictions.
- `fit(n_jobs_cv, store_predictions, ...)` - Estimate DoubleML models.
- `get_params(learner)` - Get hyperparameters for the nuisance model of DoubleML models.
- `p_adjust(method)` - Multiple testing adjustment for DoubleML models.
- `sensitivity_analysis(cf_y, cf_d, rho, ...)` - Performs a sensitivity analysis to account for unobserved confounders.
- `sensitivity_benchmark(benchmarking_set, ...)` - Computes a benchmark for a given set of features.
- `sensitivity_plot(idx_treatment, value, ...)` - Contour plot of the sensivity with respect to latent/confounding variables.
- `set_ml_nuisance_params(learner, treat_var, ...)` - Set hyperparameters for the nuisance models of DoubleML models.
- `set_sample_splitting(all_smpls, ...)` - Set the sample splitting for DoubleML models.
- `tune(param_grids, tune_on_folds, ...)` - Hyperparameter-tuning for DoubleML models.
- `tune_ml_models(ml_param_space, ...)` - Hyperparameter-tuning for DoubleML models using Optuna.
```

--------------------------------

### Import libraries for DoubleML pension QTE example

Source: https://docs.doubleml.org/dev/examples/py_double_ml_pension_qte.html

Imports required Python libraries including numpy, pandas, doubleml, sklearn, lightgbm, and plotting libraries. This is the initial setup for the example.

```python
import numpy as np
import pandas as pd
import doubleml as dml
import multiprocessing
from doubleml.datasets import fetch_401K

from sklearn.base import clone

from lightgbm import LGBMClassifier, LGBMRegressor

import matplotlib.pyplot as plt
import seaborn as sns
```

--------------------------------

### Install Python prerequisites on Linux

Source: https://docs.doubleml.org/dev/intro/install.html

Install python3 and python3-pip using the system package manager on Linux.

```bash
$ sudo apt-get install python3 python3-pip
```

--------------------------------

### DoubleMLAPOS Tuning Example

Source: https://docs.doubleml.org/dev/api/generated/doubleml.irm.DoubleMLAPOS.html

Example demonstrating how to tune DoubleML models using Optuna with DoubleMLAPOS.

```APIDOC
## DoubleMLAPOS Tuning Example

### Description
This example shows how to tune a DoubleML model using Optuna, including defining parameter grid functions, setting up the Optuna sampler, and fitting the model.

### Usage
```python
import numpy as np
from doubleml import DoubleMLData, DoubleMLPLR
from doubleml.plm.datasets import make_plr_CCDDHNR2018
from lightgbm import LGBMRegressor
import optuna

# Generate data
np.random.seed(42)
data = make_plr_CCDDHNR2018(n_obs=500, dim_x=20, return_type='DataFrame')
dml_data = DoubleMLData(data, 'y', 'd')

# Initialize model
dml_plr = DoubleMLPLR(
    dml_data,
    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42),
    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42)
)

# Define parameter grid functions
def ml_l_params(trial):
    return {
        'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
    }

def ml_m_params(trial):
    return {
        'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
    }

ml_param_space = {'ml_l': ml_l_params, 'ml_m': ml_m_params}

# Tune with TPE sampler
optuna_settings = {
    'n_trials': 5,
    'sampler': optuna.samplers.TPESampler(seed=42),
}
tune_res = dml_plr.tune_ml_models(ml_param_space, optuna_settings=optuna_settings, return_tune_res=True)
print(tune_res[0]['ml_l'].best_params)
# {'learning_rate': 0.03907122389107094}

# Fit and get results
dml_plr.fit().summary
#       coef   std err          t         P>|t|     2.5 %    97.5 %
# d  0.57436  0.045206  12.705519  5.510257e-37  0.485759  0.662961
```

### Example with scoring methods and directions
```python
scoring_methods = {
    'ml_l': 'neg_mean_squared_error',  # Negative metric
    'ml_m': 'neg_mean_squared_error'
}
optuna_settings = {
    'n_trials': 50,
    'direction': 'maximize',  # Maximize negative MSE (minimize MSE)
    'sampler': optuna.samplers.TPESampler(seed=42),
}
tune_res = dml_plr.tune_ml_models(ml_param_space, scoring_methods=scoring_methods,
                                  optuna_settings=optuna_settings, return_tune_res=True)
print(tune_res[0]['ml_l'].best_params)
# {'learning_rate': 0.04300012336462904}
dml_plr.fit().summary
#       coef   std err          t         P>|t|     2.5 %    97.5 %
# d  0.574796  0.045062  12.755721  2.896820e-37  0.486476  0.663115
```

```

--------------------------------

### DoubleML Setup and Estimation

Source: https://docs.doubleml.org/stable/examples/py_double_ml_rdflex.ipynb

This snippet shows the typical workflow for DoubleML: importing the library, creating a DoubleMLData object, and fitting a DoubleMLPLR model.

```python
import doubleml as dml
from doubleml import DoubleMLData
from doubleml import DoubleMLPLR

dml_data = DoubleMLData(df, y_col='y', d_cols='d', x_cols='x')

dml_plr = DoubleMLPLR(dml_data, ml_g, ml_m)
dml_plr.fit()
```

--------------------------------

### Import libraries for RDFlex RDD example

Source: https://docs.doubleml.org/dev/examples/py_double_ml_rdflex.html

Imports required packages for the RDFlex example, including numpy, pandas, plotly, statsmodels, lightgbm, rdrobust, and doubleml.rdd. The dependencies for doubleml.rdd can be installed separately via pip install rdrobust if necessary.

```python
import numpy as np
import pandas as pd

import plotly.express as px
import plotly.graph_objs as go
from statsmodels.nonparametric.kernel_regression import KernelReg

from lightgbm import LGBMRegressor, LGBMClassifier

from rdrobust import rdrobust

import doubleml as dml
from doubleml.rdd import RDFlex
from doubleml.rdd.datasets import make_simple_rdd_data
```

--------------------------------

### Set up and activate virtualenv on Linux

Source: https://docs.doubleml.org/dev/intro/install.html

Create a virtual environment named dml-venv and activate it to avoid package conflicts.

```bash
$ virtualenv -p python3 dml-venv
$ source dml-venv/bin/activate
```

--------------------------------

### DoubleMLPLIV Usage Example

Source: https://docs.doubleml.org/dev/api/generated/doubleml.plm.DoubleMLPLIV.html

Example demonstrating how to use DoubleMLPLIV for partially linear IV regression, including data generation, model setup, tuning, and fitting.

```APIDOC
## DoubleMLPLIV Usage Example

### Description
This example shows how to use the DoubleMLPLIV class for partially linear IV regression. It covers data generation, model initialization, hyperparameter tuning with Optuna, and fitting the model to obtain estimates.

### Class
DoubleMLPLIV

### Module
doubleml.plm

### Usage Example
```python
>>> import numpy as np
>>> from doubleml import DoubleMLData, DoubleMLPLR
>>> from doubleml.plm.datasets import make_plr_CCDDHNR2018
>>> from lightgbm import LGBMRegressor
>>> import optuna
>>> # Generate data
>>> np.random.seed(42)
>>> data = make_plr_CCDDHNR2018(n_obs=500, dim_x=20, return_type='DataFrame')
>>> dml_data = DoubleMLData(data, 'y', 'd')
>>> # Initialize model
>>> dml_plr = DoubleMLPLR(
...    dml_data,
...    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42),
...    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42)
... )
>>> # Define parameter grid functions
>>> def ml_l_params(trial):
...     return {
...         'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
...     }
>>> def ml_m_params(trial):
...     return {
...         'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
...     }
>>> ml_param_space = {'ml_l': ml_l_params, 'ml_m': ml_m_params}
>>> # Tune with TPE sampler
>>> optuna_settings = {
...     'n_trials': 5,
...     'sampler': optuna.samplers.TPESampler(seed=42),
... }
>>> tune_res = dml_plr.tune_ml_models(ml_param_space, optuna_settings=optuna_settings, return_tune_res=True)
>>> print(tune_res[0]['ml_l'].best_params)  
{'learning_rate': 0.03907122389107094}
>>> # Fit and get results
>>> dml_plr.fit().summary 
      coef   std err          t         P>|t|     2.5 %    97.5 %
d  0.57436  0.045206  12.705519  5.510257e-37  0.485759  0.662961
>>> # Example with scoring methods and directions
>>> scoring_methods = {
...     'ml_l': 'neg_mean_squared_error',  # Negative metric
...     'ml_m': 'neg_mean_squared_error'
... }
>>> optuna_settings = {
...     'n_trials': 50,
...     'direction': 'maximize',  # Maximize negative MSE (minimize MSE)
...     'sampler': optuna.samplers.TPESampler(seed=42),
... }
>>> tune_res = dml_plr.tune_ml_models(ml_param_space, scoring_methods=scoring_methods,
...                                   optuna_settings=optuna_settings, return_tune_res=True)
>>> print(tune_res[0]['ml_l'].best_params)  
{'learning_rate': 0.04300012336462904}
>>> dml_plr.fit().summary 
       coef   std err          t         P>|t|     2.5 %    97.5 %
d  0.574796  0.045062  12.755721  2.896820e-37  0.486476  0.663115
```

### Methods
- **tune_ml_models(ml_param_space, optuna_settings=None, scoring_methods=None, return_tune_res=False)** - Tunes the machine learning models using Optuna. Parameters: `ml_param_space` (dict) - parameter grid functions for each nuisance model; `optuna_settings` (dict) - settings for Optuna optimization; `scoring_methods` (dict) - scoring methods for each model; `return_tune_res` (bool) - whether to return tuning results.
- **fit()** - Fits the DoubleMLPLIV model.

### Attributes
- **summary** - DataFrame with coefficient estimates, standard errors, t-statistics, p-values, and confidence intervals.
```

--------------------------------

### DoubleMLPLR Tuning and Fitting Example

Source: https://docs.doubleml.org/stable/api/generated/doubleml.irm.DoubleMLSSM.html

Example code showing how to use DoubleMLPLR with LightGBM regressors, tune hyperparameters using Optuna, and fit the model to get summary results.

```APIDOC
## DoubleMLPLR Tuning and Fitting Example

### Description
This example demonstrates the usage of the DoubleMLPLR class for double machine learning with hyperparameter tuning via Optuna. It includes data generation, model initialization, parameter grid definition, tuning with TPE sampler, and fitting the model to obtain summary statistics.

### Method
Not applicable (Python code example)

### Endpoint
Not applicable (Python code example)

### Parameters
Not applicable (Python code example)

### Request Example
```python
>>> import numpy as np
>>> from doubleml import DoubleMLData, DoubleMLPLR
>>> from doubleml.plm.datasets import make_plr_CCDDHNR2018
>>> from lightgbm import LGBMRegressor
>>> import optuna
>>> # Generate data
>>> np.random.seed(42)
>>> data = make_plr_CCDDHNR2018(n_obs=500, dim_x=20, return_type='DataFrame')
>>> dml_data = DoubleMLData(data, 'y', 'd')
>>> # Initialize model
>>> dml_plr = DoubleMLPLR(
...    dml_data,
...    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42),
...    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42)
... )
>>> # Define parameter grid functions
>>> def ml_l_params(trial):
...     return {
...         'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
...     }
>>> def ml_m_params(trial):
...     return {
...         'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
...     }
>>> ml_param_space = {'ml_l': ml_l_params, 'ml_m': ml_m_params}
>>> # Tune with TPE sampler
>>> optuna_settings = {
...     'n_trials': 5,
...     'sampler': optuna.samplers.TPESampler(seed=42),
... }
>>> tune_res = dml_plr.tune_ml_models(ml_param_space, optuna_settings=optuna_settings, return_tune_res=True)
>>> print(tune_res[0]['ml_l'].best_params)  
{'learning_rate': 0.03907122389107094}
>>> # Fit and get results
>>> dml_plr.fit().summary 
      coef   std err          t         P>|t|     2.5 %    97.5 %
d  0.57436  0.045206  12.705519  5.510257e-37  0.485759  0.662961
```

### Response
#### Success Response
Not applicable (Python code example)

#### Response Example
Not applicable (Python code example)
```

--------------------------------

### DoubleMLDIDMulti Usage Example

Source: https://docs.doubleml.org/dev/api/generated/doubleml.did.DoubleMLDIDMulti.html

Example demonstrating how to use DoubleMLDIDMulti (via DoubleMLPLR) to generate data, initialize the model, tune hyperparameters with Optuna, and fit the model to obtain results.

```APIDOC
## DoubleMLDIDMulti Usage Example

### Description
This example shows how to use the DoubleMLDIDMulti class (represented here by DoubleMLPLR) to perform double machine learning for partially linear regression models. It covers data generation, model initialization, hyperparameter tuning with Optuna, and model fitting.

### Method
Python API

### Endpoint
Not applicable (Python library)

### Parameters
- `n_obs` (int) - Number of observations (e.g., 500)
- `dim_x` (int) - Number of covariates (e.g., 20)
- `return_type` (str) - Type of data structure returned (e.g., 'DataFrame')
- `data` (DataFrame) - Input data for the model
- `y` (str) - Name of the outcome variable
- `d` (str) - Name of the treatment variable
- `ml_l` (estimator) - Machine learning estimator for the nuisance function l
- `ml_m` (estimator) - Machine learning estimator for the nuisance function m
- `ml_param_space` (dict) - Dictionary of parameter grid functions for each nuisance function
- `optuna_settings` (dict) - Settings for Optuna optimization (e.g., n_trials, sampler, direction)
- `scoring_methods` (dict) - Scoring methods for each nuisance function (e.g., 'neg_mean_squared_error')
- `return_tune_res` (bool) - Whether to return tuning results

### Request Example
```python
import numpy as np
from doubleml import DoubleMLData, DoubleMLPLR
from doubleml.plm.datasets import make_plr_CCDDHNR2018
from lightgbm import LGBMRegressor
import optuna

# Generate data
np.random.seed(42)
data = make_plr_CCDDHNR2018(n_obs=500, dim_x=20, return_type='DataFrame')
dml_data = DoubleMLData(data, 'y', 'd')

# Initialize model
dml_plr = DoubleMLPLR(
    dml_data,
    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42),
    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42)
)

# Define parameter grid functions
def ml_l_params(trial):
    return {
        'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
    }

def ml_m_params(trial):
    return {
        'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
    }

ml_param_space = {'ml_l': ml_l_params, 'ml_m': ml_m_params}

# Tune with TPE sampler
optuna_settings = {
    'n_trials': 5,
    'sampler': optuna.samplers.TPESampler(seed=42),
}
tune_res = dml_plr.tune_ml_models(ml_param_space, optuna_settings=optuna_settings, return_tune_res=True)
print(tune_res[0]['ml_l'].best_params)

# Fit and get results
dml_plr.fit().summary
```

### Response
#### Success Response
- `tune_res` (list) - List of tuning results for each nuisance function, containing best parameters.
- `summary` (DataFrame) - Summary of the fitted model including coefficient, standard error, t-statistic, p-value, and confidence intervals.

#### Response Example
```python
{'learning_rate': 0.03907122389107094}
      coef   std err          t         P>|t|     2.5 %    97.5 %
d  0.57436  0.045206  12.705519  5.510257e-37  0.485759  0.662961
```

### Additional Example with Scoring Methods and Directions
```python
scoring_methods = {
    'ml_l': 'neg_mean_squared_error',  # Negative metric
    'ml_m': 'neg_mean_squared_error'
}
optuna_settings = {
    'n_trials': 50,
    'direction': 'maximize',  # Maximize negative MSE (minimize MSE)
    'sampler': optuna.samplers.TPESampler(seed=42),
}
tune_res = dml_plr.tune_ml_models(ml_param_space, scoring_methods=scoring_methods,
                                  optuna_settings=optuna_settings, return_tune_res=True)
print(tune_res[0]['ml_l'].best_params)
dml_plr.fit().summary
```

#### Response Example
```python
{'learning_rate': 0.04300012336462904}
       coef   std err          t         P>|t|     2.5 %    97.5 %
d  0.574796  0.045062  12.755721  2.896820e-37  0.486476  0.663115
```
```

--------------------------------

### DoubleMLDIDCS Usage Example

Source: https://docs.doubleml.org/dev/api/generated/doubleml.did.DoubleMLDIDCS.html

This example shows how to use the DoubleMLDIDCS class (or similar DoubleML classes) to generate data, initialize a model, tune hyperparameters with Optuna, and fit the model to get results.

```APIDOC
## DoubleMLDIDCS Usage Example

### Description
This example demonstrates the usage of the DoubleMLDIDCS class (or similar DoubleML classes) for double machine learning. It covers data generation, model initialization, hyperparameter tuning with Optuna, and fitting the model to obtain summary statistics.

### Method
Not applicable (Python class usage)

### Endpoint
Not applicable (Python class usage)

### Parameters
Not applicable (Python class usage)

### Request Example
```python
>>> import numpy as np
>>> from doubleml import DoubleMLData, DoubleMLPLR
>>> from doubleml.plm.datasets import make_plr_CCDDHNR2018
>>> from lightgbm import LGBMRegressor
>>> import optuna
>>> # Generate data
>>> np.random.seed(42)
>>> data = make_plr_CCDDHNR2018(n_obs=500, dim_x=20, return_type='DataFrame')
>>> dml_data = DoubleMLData(data, 'y', 'd')
>>> # Initialize model
>>> dml_plr = DoubleMLPLR(
...    dml_data,
...    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42),
...    LGBMRegressor(n_estimators=50, verbose=-1, random_state=42)
... )
>>> # Define parameter grid functions
>>> def ml_l_params(trial):
...     return {
...         'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
...     }
>>> def ml_m_params(trial):
...     return {
...         'learning_rate': trial.suggest_float('learning_rate', 0.01, 0.3, log=True),
...     }
>>> ml_param_space = {'ml_l': ml_l_params, 'ml_m': ml_m_params}
>>> # Tune with TPE sampler
>>> optuna_settings = {
...     'n_trials': 5,
...     'sampler': optuna.samplers.TPESampler(seed=42),
... }
>>> tune_res = dml_plr.tune_ml_models(ml_param_space, optuna_settings=optuna_settings, return_tune_res=True)
>>> print(tune_res[0]['ml_l'].best_params)  
{'learning_rate': 0.03907122389107094}
>>> # Fit and get results
>>> dml_plr.fit().summary 
      coef   std err          t         P>|t|     2.5 %    97.5 %
d  0.57436  0.045206  12.705519  5.510257e-37  0.485759  0.662961
>>> # Example with scoring methods and directions
>>> scoring_methods = {
...     'ml_l': 'neg_mean_squared_error',  # Negative metric
...     'ml_m': 'neg_mean_squared_error'
... }
>>> optuna_settings = {
...     'n_trials': 50,
...     'direction': 'maximize',  # Maximize negative MSE (minimize MSE)
...     'sampler': optuna.samplers.TPESampler(seed=42),
... }
>>> tune_res = dml_plr.tune_ml_models(ml_param_space, scoring_methods=scoring_methods,
...                                   optuna_settings=optuna_settings, return_tune_res=True)
>>> print(tune_res[0]['ml_l'].best_params)  
{'learning_rate': 0.04300012336462904}
>>> dml_plr.fit().summary 
       coef   std err          t         P>|t|     2.5 %    97.5 %
d  0.574796  0.045062  12.755721  2.896820e-37  0.486476  0.663115
```

### Response
#### Success Response (200)
- **summary** (DataFrame) - The fitted model's summary statistics including coefficient, standard error, t-statistic, p-value, and confidence intervals.

#### Response Example
```
      coef   std err          t         P>|t|     2.5 %    97.5 %
d  0.57436  0.045206  12.705519  5.510257e-37  0.485759  0.662961
```
```

--------------------------------

### Setup and Data Generation for Learner Comparison

Source: https://docs.doubleml.org/dev/examples/learners/py_learner.html

Imports DoubleML and numpy, defines simulation parameters (theta=0, n_obs=500, dim_x=5, n_rep=200), sets a random seed, and generates 200 datasets using make_irm_data with R2_d=0.8 and R2_y=0.8. Each dataset is wrapped in a DoubleMLData object with outcome 'y' and treatment 'd'. Warnings about non-positive nu2 are filtered.

```python
import numpy as np
import doubleml as dml

from doubleml.irm.datasets import make_irm_data

import warnings

warnings.filterwarnings("ignore", message="The estimated nu2 for d is not positive")

theta = 0
n_obs = 500
dim_x = 5
n_rep = 200

np.random.seed(42)
datasets = []
for i in range(n_rep):
    data = make_irm_data(theta=theta, n_obs=n_obs, dim_x=dim_x,
                         R2_d=0.8, R2_y=0.8, return_type='DataFrame')
    datasets.append(dml.DoubleMLData(data, 'y', 'd'))
```

--------------------------------

### Create and fit DoubleMLIRM model

Source: https://docs.doubleml.org/stable/examples/py_double_ml_plm_irm_hetfx.ipynb

Initializes a DoubleMLIRM object with data, treatment, outcome, and confounders, then fits the model. This is the core setup for the IRM example.

```python
aipw_obj_2 = DoubleMLIRM(data, 'd2', 'y', ['x1', 'x2', 'x3', 'x4', 'x5', 'x6', 'x7', 'x8', 'x9', 'x10', 'x11', 'x12', 'x13', 'x14', 'x15', 'x16', 'x17', 'x18', 'x19', 'x20'])
aipw_obj_2.fit()
```

--------------------------------

### Windows: Install DoubleML with pip (with virtualenv)

Source: https://docs.doubleml.org/stable/intro/install.html

Create and activate a virtual environment named 'dml-venv', then install DoubleML with pip. Verify with pip show, pip freeze, and a Python version check.

```bash
$ python -m venv dml-venv
$ dml-venv\Scripts\activate
```

```bash
$ pip install -U DoubleML
```

```bash
$ python -m pip show DoubleML # to see which version and where DoubleML is installed
$ python -m pip freeze # to see all packages installed in the active virtualenv
$ python -c "import doubleml as dml; print(dml.__version__)"
```

--------------------------------

### Load DoubleML and mlr3learners libraries

Source: https://docs.doubleml.org/dev/examples/R_double_ml_basic_iv.html

Loads the DoubleML package and the mlr3learners package, sets a random seed, and suppresses warnings. This is the initial setup for the IV example.

```R
library(DoubleML)
library(mlr3learners)

set.seed(1234)
options(warn=-1)
```

--------------------------------

### Import libraries and generate discrete treatment data

Source: https://docs.doubleml.org/dev/examples/py_double_ml_apo.html

Imports required Python libraries (numpy, pandas, matplotlib, seaborn, scikit-learn, doubleml) and the data generation function make_irm_data_discrete_treatments. This is the initial setup for the APO estimation example.

```python
import numpy as np
import pandas as pd

import matplotlib.pyplot as plt
import seaborn as sns

from sklearn.linear_model import LogisticRegression, LinearRegression
from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor

import doubleml as dml
from doubleml.irm.datasets import make_irm_data_discrete_treatments
```

--------------------------------

### DoubleMLQTE Class Usage

Source: https://docs.doubleml.org/stable/api/generated/doubleml.irm.DoubleMLQTE.html

Example demonstrating how to instantiate and fit a DoubleMLQTE model using DoubleML.

```APIDOC
## DoubleMLQTE Class

### Description
DoubleMLQTE is a class for estimating quantile treatment effects using double machine learning.

### Usage Example
```python
>>> import numpy as np
>>> import doubleml as dml
>>> from doubleml.irm.datasets import make_irm_data
>>> from sklearn.ensemble import RandomForestClassifier
>>> np.random.seed(3141)
>>> ml_g = RandomForestClassifier(n_estimators=100, max_features=20, max_depth=10, min_samples_leaf=2)
>>> ml_m = RandomForestClassifier(n_estimators=100, max_features=20, max_depth=10, min_samples_leaf=2)
>>> data = make_irm_data(theta=0.5, n_obs=500, dim_x=20, return_type='DataFrame')
>>> obj_dml_data = dml.DoubleMLData(data, 'y', 'd')
>>> dml_qte_obj = dml.DoubleMLQTE(obj_dml_data, ml_g, ml_m, quantiles=[0.25, 0.5, 0.75])
>>> dml_qte_obj.fit().summary
          coef   std err         t     P>|t|     2.5 %    97.5 %
0.25  0.274825  0.347310  0.791297  0.428771 -0.405890  0.955541
0.50  0.449150  0.192539  2.332782  0.019660  0.071782  0.826519
0.75  0.709606  0.193308  3.670867  0.000242  0.330731  1.088482
```

### Methods
- **bootstrap**([method, n_rep_boot]) - Multiplier bootstrap for DoubleML models.
- **confint**([joint, level]) - Confidence intervals for DoubleML models.
- **draw_sample_splitting**() - Draw sample splitting for DoubleML models.
- **fit**([n_jobs_models, n_jobs_cv, ...]) - Estimate DoubleMLQTE models.
- **p_adjust**([method]) - Multiple testing adjustment for DoubleML models.
- **set_sample_splitting**(all_smpls[, ...]) - Set the sample splitting for DoubleML models.
- **tune_ml_models**(ml_param_space[, ...]) - Hyperparameter-tuning for DoubleML models using Optuna.
```

--------------------------------

### Install DoubleML from a .whl file

Source: https://docs.doubleml.org/dev/intro/install.html

Installs a released version of DoubleML from a .whl file obtained from GitHub Releases. Requires python and pip to be set up.

```bash
$ pip install -U DoubleML-0.5.1-py3-none-any.whl
```

--------------------------------

### Install DoubleML development version from GitHub

Source: https://docs.doubleml.org/dev/intro/install.html

Installs the development version of DoubleML from GitHub using remotes::install_github. Requires the remotes package to be previously installed. After installation, load the package with library(DoubleML).

```R
remotes::install_github("DoubleML/doubleml-for-r")
```

```R
library(DoubleML)
```
Tokens:
Raw