Ruby Linear Regression

An implementation of a linear regression machine learning algorithm implemented in Ruby.

Features:

• Supports simple problems with one independent variable to predict a dependent variable and multivariate problems with multiple independent variables to predict a dependent variable.
• Supports training using the normal equation
• Supports training using gradient descent
• The library is implemented in pure ruby using Ruby's Matrix implementation.

Installation

  gem install ruby_linear_regression

Example of usage - Train model using normal equation

require 'matrix'
require 'csv'
require 'ruby_linear_regression'

x_data = []
y_data = []
# Load data from CSV file into two arrays - one for independent variables X (x_data) and one for the dependent variable y (y_data)
CSV.foreach("staten-island-single-family-home-sales-2015.csv", :headers => true) do |row|
  # Each row contains square feet for property and living area like this: [ SQ FEET PROPERTY, SQ FEET HOUSE ]  
  x_data.push( [row[0].to_i, row[1].to_i] )
  y_data.push( row[2].to_i )
end

# Create regression model
linear_regression = RubyLinearRegression.new

# Load training data
linear_regression.load_training_data(x_data, y_data)

# Train the model using the normal equation
linear_regression.train_normal_equation

# Output the cost
puts "Trained model with the following cost fit #{linear_regression.compute_cost}"

# Predict the price of a 2000 sq feet property with a 1500 sq feet house
prediction_data = [2000, 1500]
predicted_price = linear_regression.predict(prediction_data)
puts "Predicted selling price for a 1500 sq feet house on a 2000 sq feet property: #{predicted_price.round}$"

More details about this example implementation can be found in this blog post.

Example of usage - Train model using gradient descent

require 'matrix'
require 'csv'
require 'ruby_linear_regression'

x_data = []
y_data = []
# Load data from CSV file into two arrays - one for independent variables X (x_data) and one for the dependent variable y (y_data)
CSV.foreach("staten-island-single-family-home-sales-2015.csv", :headers => true) do |row|
  # Each row contains square feet for property and living area like this: [ SQ FEET PROPERTY, SQ FEET HOUSE ]  
  x_data.push( [row[0].to_i, row[1].to_i] )
  y_data.push( row[2].to_i )
end

# Create regression model
linear_regression = RubyLinearRegression.new

# Load training data
linear_regression.load_training_data(x_data, y_data)

# Train the model using gradient descent
linear_regression.train_gradient_descent(0.0005, 1000, true)

# Output the cost
puts "Trained model with the following cost fit #{linear_regression.compute_cost}"

# Predict the price of a 2000 sq feet property with a 1500 sq feet house
prediction_data = [2000, 1500]
predicted_price = linear_regression.predict(prediction_data)
puts "Predicted selling price for a 1500 sq feet house on a 2000 sq feet property: #{predicted_price.round}$"