sm-algo 0.1.1

Small pacakge with various algorithms

SmartAlgo: A Lightweight Machine Learning Library

SmartAlgo is a Python package providing essential machine learning algorithms implemented from scratch. Designed for simplicity and educational purposes, it includes linear regression, k-means clustering, and ridge logistic regression. Built with minimal dependencies (numpy and scipy), it's perfect for learning the fundamentals of ML algorithms.

Installation

Install the package via pip:

pip install sm_algo

Dependencies:

• numpy
• scipy

Algorithms

1. Linear Regression with Gradient Descent

Class: LinearRegression
A linear regression model trained using gradient descent.

Parameters:

• learning_rate (float, default=0.01): Step size for gradient descent.
• epochs (int, default=1000): Number of training iterations.

Methods:

• fit(X, y): Trains the model on data X and targets y.
• predict(X): Returns predicted values for input X.

Example:

from sm_algo.linreg import LinearRegression
from sklearn.datasets import load_diabetes
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler

# Load data
data = load_diabetes()
X, y = data.data, data.target
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Normalize data
scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

# Train model
model = LinearRegression(learning_rate=0.1, epochs=1000)
model.fit(X_train, y_train)

# Predict and evaluate
y_pred = model.predict(X_test)
print(f"MSE: {mean_squared_error(y_test, y_pred):.2f}")
print(f"R²: {r2_score(y_test, y_pred):.2f}")

2. K-Means Clustering

Class: KMeans
K-Means clustering with k-means++ centroid initialization.

Parameters:

• n_clusters (int, default=8): Number of clusters.
• max_iter (int, default=300): Maximum iterations per run.
• tol (float, default=1e-4): Tolerance to declare convergence.
• random_state (int, optional): Seed for random centroid initialization.

Methods:

• fit(X): Computes clustering on X.
• predict(X): Predicts cluster indices for new data.

Example:

from sm_algo.kmeans import KMeans
from sklearn.datasets import load_iris
import matplotlib.pyplot as plt

# Load data
iris = load_iris()
X = iris.data

# Cluster data
kmeans = KMeans(n_clusters=3)
kmeans.fit(X)
labels = kmeans.labels_

# Visualize clusters (first two features)
plt.scatter(X[:, 0], X[:, 1], c=labels, cmap='viridis')
plt.scatter(kmeans.centroids[:, 0], kmeans.centroids[:, 1], marker='X', s=200, c='red')
plt.xlabel('Sepal Length')
plt.ylabel('Sepal Width')
plt.show()

3. Ridge Logistic Regression

Class: LogisticRegressionRidge
Logistic regression with L2 regularization, trained via gradient descent.

Parameters:

• learning_rate (float, default=0.01): Gradient descent step size.
• lambda_ (float, default=0.1): L2 regularization strength.
• epochs (int, default=1000): Training iterations.
• fit_intercept (bool, default=True): Whether to add a bias term.
• verbose (bool, default=False): Print training loss every 100 epochs.

Methods:

• fit(X, y): Trains the model.
• predict_proba(X): Returns class probabilities.
• predict(X, threshold=0.5): Returns class labels.

Example:

from sm_algo.logisticreg import LogisticRegressionRidge
from sklearn.datasets import make_classification
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt

# Generate synthetic data
X, y = make_classification(n_samples=1000, n_features=2, n_classes=2, random_state=42)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Train model
model = LogisticRegressionRidge(learning_rate=0.001, lambda_=0.1, epochs=1000)
model.fit(X_train, y_train)

# Predict
y_pred = model.predict(X_test)
print(f"Accuracy: {accuracy_score(y_test, y_pred):.2f}")

# Optional: Plot decision boundary (requires helper functions)
def plot_decision_boundary(model, X, y):
    # Implementation from test example
    pass

plot_decision_boundary(model, X_test, y_test)

Contributing

Contributions are welcome! Please submit issues or pull requests on the GitHub repository.