@geoda/regression

This package provides implementations of various spatial regression models for analyzing spatial data. It includes the following models:

Features

• OLS Regression
• Spatial Lag Model
• Spatial Error Model

Installation

yarn add @geoda/regression

Available Models

1. OLS Regression

Basic Ordinary Least Squares regression for spatial data analysis.

2. Spatial Lag Model

A spatial regression model that incorporates spatial dependence in the dependent variable. The model includes:

• Spatial lag coefficient (Rho)
• Maximum likelihood estimation
• Comprehensive diagnostics including:
  • Heteroskedasticity tests (Breusch-Pagan)
  • Spatial dependence tests (Likelihood Ratio)
  • Model fit statistics (R-squared, AIC, etc.)

3. Spatial Error Model

A spatial regression model that accounts for spatial autocorrelation in the error terms. The model includes:

• Spatial error coefficient (Lambda)
• Maximum likelihood estimation
• Comprehensive diagnostics including:
  • Heteroskedasticity tests (Breusch-Pagan)
  • Spatial dependence tests (Likelihood Ratio)
  • Model fit statistics (R-squared, AIC, etc.)

Usage

Each model provides detailed output including:

• Model coefficients and standard errors
• Model fit statistics
• Diagnostic tests for spatial dependence and heteroskedasticity
• Variable-specific statistics

Example

import { spatialLagRegression } from '@geoda/regression';

const result = await spatialLagRegression({
  x: independentVariables,
  y: dependentVariable,
  weights: spatialWeights,
  xNames: ['var1', 'var2'],
  yName: 'target',
  datasetName: 'myDataset'
});

Output Format

The regression results include:

• Basic model information (dataset name, number of observations, etc.)
• Model coefficients and their significance
• Model fit statistics (R-squared, AIC, etc.)
• Diagnostic tests for spatial effects
• Detailed variable-specific statistics

For more detailed information about each model, please refer to the Spatial Regression documentation.