BuildML

BuildML is a Python machine-learning library designed to simplify the process of data preparation, feature engineering, model building, and evaluation. It provides a collection of tools for both classification and regression tasks, as well as functionalities for data exploration and manipulation. BuildML is a distribution of the TechLeo community to make complex machine-learning processes, easy.

Author

TechLeo

• Email: techleo.ng@outlook.com
• GitHub: TechLeo GitHub
• LinkedIn: TechLeo LinkedIn

Contact

For inquiries, suggestions, or feedback, please feel free to reach out to the author:

• Team Lead: Onyiriuba Leonard
• Personal Email: workwithtechleo@gmail.com
• Company Email: techleo.ng@outlook.com
• GitHub Issues: BuildML Issues
• LinkedIn Messages: TechLeo LinkedIn

Your feedback is valuable and contributes to the continuous improvement of BuildML. The author welcomes collaboration and looks forward to hearing from the users of MLwiz.

Features

Features from the current release.

Data Loading and Handling

• get_dataset: Load a dataset.
• get_training_test_data: Split the dataset into training and test sets.
• load_large_dataset: Load a large dataset efficiently.
• reduce_data_memory_useage: Reduce memory usage of the dataset.

Data Cleaning and Manipulation

• drop_columns: Drop specified columns from the dataset.
• fix_missing_values: Handle missing values in the dataset.
• fix_unbalanced_dataset: Address class imbalance in a classification dataset.
• filter_data: Filter data based on specified conditions.
• remove_duplicates: Remove duplicate rows from the dataset.
• rename_columns: Rename columns in the dataset.
• replace_values: Replace specified values in the dataset.
• reset_index: Reset the index of the dataset.
• set_index: Set a specific column as the index.
• sort_index: Sort the index of the dataset.
• sort_values: Sort the values of the dataset.

Data Formatting and Transformation

• categorical_to_datetime: Convert categorical columns to datetime format.
• categorical_to_numerical: Convert categorical columns to numerical format.
• numerical_to_categorical: Convert numerical columns to categorical format.
• column_binning: Bin values in a column into specified bins.

Exploratory Data Analysis

• eda: Perform exploratory data analysis on the dataset.
• eda_visual: Visualize exploratory data analysis results.
• pandas_profiling: Generate a Pandas Profiling report for the dataset.
• sweetviz_profile_report: Generate a Sweetviz Profile Report for the dataset.
• count_column_categories: Count the categories in a categorical column.
• unique_elements_in_columns: Get the unique elements that exist in each column in the dataset.

Feature Engineering

• extract_date_features: Extract date-related features from a datetime column.
• polyreg_x: Get the polynomial regression x for independent variables after specifying the degree.
• select_features: Select relevant features for modeling.
• select_dependent_and_independent: Select dependent and independent variables.

Data Preprocessing

• scale_independent_variables: Scale independent variables in the dataset.
• remove_outlier: Remove outliers from the dataset.
• split_data: Split the dataset into training and test sets.

Model Building and Evaluation

• poly_get_optimal_degree: Find the best degree for polynomial regression.
• get_bestK_KNNregressor: Find the best K value for KNN regression.
• train_model_regressor: Train a regression model.
• regressor_predict: Make predictions using a regression model.
• regressor_evaluation: Evaluate the performance of a regression model.
• regressor_model_testing: Test a regression model.
• polyreg_graph: Visualize a polynomial regression graph.
• simple_linregres_graph: Visualize a regression graph.
• build_multiple_regressors: Build multiple regression models.
• build_multiple_regressors_from_features: Build regression models using selected features.
• build_single_regressor_from_features: Build a single regression model using selected features.
• get_bestK_KNNclassifier: Find the best K value for KNN classification.
• train_model_classifier: Train a classification model.
• classifier_predict: Make predictions using a classification model.
• classifier_evaluation: Evaluate the performance of a classification model.
• classifier_model_testing: Test a classification model.
• classifier_graph: Visualize a classification graph.
• build_multiple_classifiers: Build multiple classification models.
• build_multiple_classifiers_from_features: Build classification models using selected features.
• build_single_classifier_from_features: Build a single classification model using selected features.

Data Aggregation and Summarization

• group_data: Group and summarize data based on specified conditions.

Data Type Handling

• select_datatype: Select columns of a specific datatype in the dataset.

Installation

You can install BuildML using pip:

pip install buildml

Example Usage

Example 1

import numpy as np
import pandas as pd
from sklearn.linear_model import LogisticRegression
from sklearn.ensemble import RandomForestClassifier, DecisionTreeClassifier
from sklearn.snm import SVC
from buildml import SupervisedLearning


dataset = pd.read_csv("Your_file_path")  # Load your dataset(e.g Pandas DataFrame)
data = SupervisedLearning(dataset)

# Exploratory Data Analysis
eda = data.eda()
eda_visual = data.eda_visual()

# Build and Evaluate Classifier
classifiers = [
    "LogisticRegression(random_state = 0)", 
    "RandomForestClassifier(random_state = 0)", 
    "DecisionTreeClassifier(random_state = 0)", 
    "SVC()"
    ]
    
build_model = data.build_multiple_classifiers(classifiers, 
                                          kfold=5, 
                                          cross_validation=True, 
                                          graph=True, 
                                          length=8, 
                                          width=12)

Example 2: Working on a dataset with train and test data given.

# Import Libraries
import pandas as pd
import numpy as np
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.ensemble import RandomForestClassifier
from sklearn.tree import DecisionTreeClassifier
from xgboost import XGBClassifier
from buildml import SupervisedLearning

# Get Dataset
training_data = pd.read_csv("train.csv")
test_data = pd.read_csv("test.csv")

dataset = pd.concat([training_data, test_data], axis = 0)

# BuildML on Dataset
automate_training = SupervisedLearning(training_data)
automate_test = SupervisedLearning(test_data)

automate = [automate_training, automate_test]

# Exploratory Data Analysis
training_eda = automate_training.eda()
test_eda = automate_test.eda()

# Data Cleaning and Transformation 
training_eda_visual = automate_training.eda_visual(
                            figsize_barchart = (55, 10), 
                            figsize_heatmap = (15, 10), 
                            figsize_histogram=(35, 20)
                            )

for data in automate:
    data.reduce_data_memory_useage()
    data.drop_columns("Drop irrelevant columns")
    data.categorical_to_numerical() # If your data has categorical features

select_variables = automate_training.select_dependent_and_independent(predict = "Loan Status")

# Further Data Preparation and Segregation
training_data_clean = automate_training.get_dataset()
test_data_clean = automate_test.get_dataset()

unbalanced_dataset_check = automate_training.count_column_categories(column = "Specify what you are predicting")
split_data = automate_training.split_data()
fix_unbalanced_data = automate_training.fix_unbalanced_dataset(
                                        sampler = "RandomOverSampler", 
                                        random_state = 0
                                        )

check_unbalanced_data_fix = automate_training.count_column_categories(
                                    column = "Specify what you are predicting", 
                                    test_data = True
                                    )

# Model Building 
classifiers = [LogisticRegression(random_state = 0),
                SVC(),
                RandomForestClassifier(random_state = 0),
                DecisionTreeClassifier(random_state = 0),
                XGBClassifier(random_state = 0)
                ]

build_model = automate_training.build_multiple_classifiers(classifiers = classifiers,
                                                            kfold = 10,
                                                            cross_validation = True,
                                                            graph = True
                                                            )

Acknowledgments

BuildML relies on several open-source libraries to provide its functionality. We would like to express our gratitude to the developers and contributors of the following libraries:

• NumPy
• Pandas
• Matplotlib
• Seaborn
• yData Profiling
• Sweetviz
• Imbalanced-Learn (imblearn)
• Scikit-learn
• Warnings
• Datatable

The BuildML library builds upon the functionality provided by these excellent tools, We sincerely thank the maintainers and contributors of these libraries for their valuable contributions to the open-source community.

❤️ Support BuildML

If you find BuildML helpful and would like to support its development, there are several ways you can contribute:

☕ Buy Me a Coffee

You can show your appreciation by buying me a coffee. Every little bit helps and goes directly towards keeping this project maintained and improving.

🌟 Sponsorship

Consider becoming a sponsor to provide ongoing support. Sponsors receive special recognition and exclusive perks:

• 🎉 Exclusive updates on the latest features and developments.
• 🚀 Early access to pre-releases and beta versions.
• 📢 Special mention in the project documentation and README.

Become a sponsor by clicking on the "Sponsor" button on the BuildML repository.

👩‍💻 Contribute

If you're a developer, you can contribute directly by:

• 🐛 Reporting bugs or suggesting improvements by opening issues.
• 🛠 Submitting pull requests to enhance the codebase.

No contribution is too small, and your involvement is highly appreciated!

Thank you for considering supporting BuildML! Your generosity keeps the project alive and thriving. 🚀

License

BuildML is distributed under the MIT License. Feel free to use, modify, and distribute it according to the terms of the license.

Changelog

v1.0.0 (January 2024):

• First release

v1.0.1 (January 2024):

• Removed boxplot graph from eda_visual.
• Added new parameters for eda_visual.
• Removed user_guide from SupervisedLearning parameter in documentation.
• Improved documentation.

v1.0.2 (January 2024):

• Improved documentation.
• Fix for building all models without splitting the data.
• Updated requirements.txt file.
• Fixed dependency error in installation.

v1.0.3 (January 2024):

• Allow the method count_column_categories to work for split_data.
• Improved documentation.

v1.0.4 (January 2024):

• Improved documentation.

v1.0.5 (January 2024):

• Improved documentation.
• Created external documentation for BuildML.

v1.0.6 (January 2024):

• Fixed polynomial regression graph
• Updated documentation on build_single_regressor_from_features
• Updated documentation on build_single_classifier_from_features
• Imporved documentation for polyreg_graph
• Imporved documentation for split_data
• Added option to specify the test_size in split_data but default parameter remains 0.2
• Created method poly_get_optimal_degree to find the best degree for polynomial regression.
• Updated README.md file.

v1.0.7 (January 2024):

• Fixed self.__polynomial_regression not defined.
• Improved documentation for build_single_classifier_from_features.
• Improved documentation for build_single_regressor_from_features.
• Improved documentation for build_multiple_classifier_from_features.
• Improved documentation for build_multiple_regressor_from_features.

v1.0.8 (January 2024):

• Fixed unstable release v1.0.7.

v1.0.9 (January 2024):

• Fixed unstable release v1.0.8.

Contributors

We'd like to express our gratitude to the following contributors who have influenced and supported BuildML:

• Onyiriuba Leonard: for overseeing the entire project development lifecycle.
• Role: Project Lead and Maintainer.
• Email: workwithtechleo@gmail.com.

• The TechLeo Community: for allowing the use of this project as a way to explain, learn, test, understand, and make easy, the machine learning process.
• Role: Testers.
• Email: techleo.ng@gmail.com.