Regression models

Simple demonstration of the different implemented regression models in emlearn

import os.path

import emlearn
import numpy
import pandas
import matplotlib.pyplot as plt
import seaborn

try:
    # When executed as regular .py script
    here = os.path.dirname(__file__)
except NameError:
    # When executed as Jupyter notebook / Sphinx Gallery
    here = os.getcwd()

Create dataset

Using a simple multi-class dataset included with scikit-learn

def load_dataset():
    from sklearn import datasets
    data = datasets.load_diabetes(as_frame=True)

    df = data.data.copy()
    df.columns = data.feature_names
    df['target'] = data.target

    return df

dataset = load_dataset()

def plot_results(ax, model, X, y):
    from sklearn.metrics import PredictionErrorDisplay

    disp = PredictionErrorDisplay.from_estimator(
        model, X, y,
        ax=ax,
        kind="actual_vs_predicted",
    )

Train, convert and run model

Using the standard scikit-learn process, and then using emlearn to convert the model to C

def build_run_regressor(model, name, ax=None, feature_columns=None, color=None):
    from sklearn.model_selection import train_test_split

    target_column = 'target'
    if feature_columns is None:
        feature_columns = list(set(dataset.columns) - set([target_column]))

    # Train model
    test, train = train_test_split(dataset, test_size=0.3, random_state=3)

    model.fit(train[feature_columns], train[target_column])

    out_dir = os.path.join(here, 'regression')
    if not os.path.exists(out_dir):
        os.makedirs(out_dir)

    model_filename = os.path.join(out_dir, f'{name}_model.h')
    cmodel = emlearn.convert(model, method='loadable')
    code = cmodel.save(file=model_filename, name='model')

    test_pred = cmodel.predict(test[feature_columns])
    # Generate a test dataet
    test_data = numpy.array(test[feature_columns]).flatten()
    test_res = numpy.array(test_pred).flatten()
    # TODO: use tools to generate C file to compare classifier against expected results i C

    if ax is not None:
        plot_results(ax, model, X=test[feature_columns], y=test[target_column])

Run all models

Some of the supported modela and configurations

import sklearn.ensemble
import sklearn.tree
import sklearn.neural_network
import sklearn.naive_bayes

models = {
    'random_forest': sklearn.ensemble.RandomForestRegressor(n_estimators=10, random_state=1),
    'extra_trees': sklearn.ensemble.ExtraTreesRegressor(n_estimators=10, random_state=1),
    'decision_tree': sklearn.tree.DecisionTreeRegressor(),

    # TODO: support MultiLayerPerceptron
    #'sklearn_mlp': sklearn.neural_network.MLPClassifier(hidden_layer_sizes=(10,10,), max_iter=30, random_state=1),

    # TODO: support the various LinearRegression alternatives
}

# Based the below feature analysis - s5, bmi, bp are the most useful features for diabetes dataset
# https://scikit-learn.org/stable/auto_examples/ensemble/plot_gradient_boosting_regression.html#gsphx-glr-auto-examples-ensemble-plot-gradient-boosting-regression-py
feature_columns = ['s5', 'bmi', 'bp']

fig, axs = plt.subplots(
        #ncols=len(models),
        nrows=len(models),
        figsize=(4, 4.2*len(models)),
        sharex=False, sharey=False,
)
for model_no, (name, cls) in enumerate(models.items()):
    ax = axs[model_no]
    build_run_regressor(cls, name, ax)

    ax.set_title(name)
    ax.set_xlabel('')

plt.show()