EasyLM

EasyLM is a Python library for linear regression with R-style summaries, automatic coefficient interpretation, and model comparison. It provides simple interfaces, comprehensive statistics, and built-in visualizations.

---

Features

• R-like lm() interface with comprehensive summaries
• Coefficients, standard errors, t-values, p-values, R², AIC, BIC
• Compare multiple models side-by-side
• Built-in plotting for model comparison
• Accepts NumPy arrays, Pandas DataFrames, and Python lists
• Clean, modular, and extensible architecture

---

Installation

Install directly from PyPI:

Installation

You can install from TestPyPI:

pip install amoang-easylm

Or install the latest version from GitHub:

git clone https://github.com/yourusername/EasyLM.git
cd EasyLM
pip install -r requirements.txt
pip install -e .

Basic Usage

from EasyLM import LinearModel
import numpy as np

# Create sample data
X = np.array([[1], [2], [3], [4], [5]])
y = np.array([2, 4, 5, 4, 5])

# Fit model
model = LinearModel()
model.fit(X, y)

# View R-style summary
print(model.summary())

Example Output

After fitting a simple model, you’ll see an R‑style summary:

Call: EasyLM LinearModel

Observations: 5
Parameters: 2
Degrees of Freedom (resid): 3

Coefficients:
   Coef.  Std.Err.  t value  Pr(>|t|)
0  1.5000    0.7906   1.8974    0.1535
1  0.7000    0.2280   3.0702    0.0547

Residual variance (sigma^2): 0.7000
R-squared: 0.7588
AIC: 10.7319
BIC: 9.9465 

Key Concepts

Linear regression fits the model:

y = β₀ + β₁x₁ + β₂x₂ + … + ε

• y: target variable
• x: features
• β: coefficients
• ε: residual error

Summary Output

• Coef.: coefficient values
• Std.Err.: standard error
• t value: coefficient ÷ Std.Err.
• Pr(>|t|): p-value (<0.05 considered significant)
• R²: variance explained
• AIC/BIC: model selection metrics
• Residual variance: remaining error

Project Structure

EasyLM/
├── EasyLM/
│   ├── __init__.py
│   ├── linear_model.py
│   ├── data_preprocessor.py
│   ├── regression_stats.py
│   ├── summary_formatter.py
│   ├── model_comparator.py
│   ├── plot_helper.py
│   └── utils.py
├── tests/
│   ├── __init__.py
│   ├── test_linear_model.py
│   ├── test_model_comparator.py
│   └── test_utils.py
├── examples/
│   └── quickstart.ipynb
├── README.md
├── requirements.txt
├── pyproject.toml
└── LICENSE

Core Classes

Classes and Their Purpose

Class	Purpose
LinearModel	Fit and predict linear models
DataPreprocessor	Validate and transform data
RegressionStats	Compute statistical quantities
SummaryFormatter	Generate R-style summaries
ModelComparator	Compare multiple models
PlotHelper	Visualization utilities

Testing

Run tests with pytest:

pip install pytest
pytest tests/ -v
pytest --cov=EasyLM tests/ 

Comparison with Alternatives

📊 Comparison with Alternatives

Feature	EasyLM	scikit-learn	statsmodels
R-style summaries	✅	❌	✅
Easy to learn	✅	✅	❌
Statistical tests	✅	❌	✅
Model comparison	✅	❌	⚠️
Lightweight	✅	❌	❌
Extensible	✅	✅	✅

When to Use

Use EasyLM if you want:

• R‑style regression output in Python
• Easy model comparison
• Teaching/learning regression concepts
• Clean, extensible code

Use alternatives when you need:

• Production ML pipelines → scikit‑learn
• Advanced econometrics → statsmodels
• Deep learning → TensorFlow/PyTorch

Contributors

Core Team

[Kyle Paolo G Bautro] – Initial work & Architecture

• Designed clean separation of concerns
• Implemented RegressionStats engine
• Created visualization tools

[Carl Dane Penano, Joshua Oraiz, Symond Ridge Fernandez] – Testing & Documentation

• Built comprehensive test suite
• Wrote user documentation
• Quality assurance

[Yochanan Bangoy] – Features & Extensions

• Refactored plotting method
• Enhanced plotting capabilities
• Performance optimizations

Special Thanks

• Inspired by R's lm() function
• Documentation style influenced by scikit-learn
• Jesus Christ

---

License

This project is licensed under the MIT License – see the LICENSE file for details.

---

Support

• Email: kyleforthedamaged@gmail.com
• Issues: GitHub Issues
• Discussions: GitHub Discussions

---

Roadmap

Version 1.0 (Current)

• Basic OLS regression
• R-style summaries
• Model comparison
• Visualization tools

Version 1.1 (Planned)

• Weighted least squares
• Ridge & Lasso regression
• Cross-validation utilities
• Polynomial features

Version 2.0 (Future)

• Generalized Linear Models (GLM)
• Time series regression
• Mixed effects models
• Interactive web dashboard