Gain insights into simplifying machine learning with PyCaret. Delve into regression, classification, clustering, and anomaly detection, and learn to deploy applications using Streamlit.

Simplifying Machine Learning with PyCaret_ .png

c.tar.gz

jupyter

RunCode

spa1

PyCaret is an low-code, open-source, machine learning library for Python. It can be used in machine learning tasks such as data preparation and model deployment.

In this course, you will learn multiple topics related to machine learning. You will start with a brief introduction to the basic concepts of machine learning, and then continue with case studies of regression, classification, clustering, and anomaly detection based on the respective modules of the PyCaret library. Finally, we will focus on using the Streamlit library to develop and deploy machine learning applications.

By the end of this course, you will have the skills to deploy the robust PyCaret library for any of your machine learning projects.

Simplifying Machine Learning with PyCaret in Python

# Creating the model
We’ll use the `create_model()` function to train the Linear Discriminant Analysis model because it performed best in the model comparison.

# Creating the model

 model = create_model('lda')

Model

align-center

This function uses **stratified $k$-fold cross-validation** to evaluate model accuracy, a variation of the standard $k$-fold technique used in the Regression chapter. The dataset is consecutively partitioned into $k$ subsamples, with one subsample being retained for validation, while the rest are used to train the model. The difference between stratified $k$-fold and standard $k$-fold is that subsamples are stratified to preserve the percentage of samples for each class, in this case, $33.3$% for every Iris species. After the stratified $k$-fold cross-validation is completed, the resulting metrics for each subsample are printed, including the mean and standard deviation values.


This function uses stratified $k$-fold cross-validation to evaluate model accuracy, a variation of the standard $k$-fold technique used in the chapter about regression. The dataset is consecutively partitioned into k subsamples. One subsample is retained for validation, while the rest are used to train the model. The difference between stratified $k$-fold and standard $k$-fold is that subsamples are stratified to preserve the percentage of samples for each class, in this case $33.3$% for every Iris species. After the stratified $k$-fold cross-validation is completed, the resulting metrics for each subsample are printed, including the mean and standard deviation values.



# Tuning the model
The Linear Discriminant Analysis model has excellent performance, with an accuracy of $98.33$%. Unfortunately, not every model performed as well. The CatBoost Classifier model achieved an accuracy of only $94.17$%. The `tune_model()` function can improve the performance of a classification model by finding the optimal combination of its hyperparameters, a technique known as **hyperparameter tuning**. By default, this function uses the [randomized search method](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.RandomizedSearchCV.html)  of the scikit-learn library, but there are various other alternatives, including the `scikit-optimize` and optuna libraries. In this case, we selected the scikit-optimize library that uses Bayesian optimization to model the hyperparameter search space and finds the optimal combination as soon as possible.

In classification tasks, learn to create, tune, plot, save and make predictions from the machine learning model.

	Accuracy	AUC	Recall	Prec.	F1	Kappa	MCC
0	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000
1	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000
2	0.9167	1.0000	0.9167	0.9333	0.9153	0.8750	0.8843
3	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000
4	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000
5	0.9167	1.0000	0.9167	0.9333	0.9153	0.8750	0.8843
6	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000
7	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000
8	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000
9	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000	1.0000
Mean	0.9833	1.0000	0.9833	0.9867	0.9831	0.9750	0.9769
SD	0.0333	0.0000	0.0333	0.0267	0.0339	0.0500	0.0463

Introduction to Machine Learning

Regression

Classification

Clustering

Anomaly Detection

Natural Language Processing

Deploying a Machine Learning Model

Conclusion

Appendix

Building the Model

Creating the model

Model