Statistics > Correlation and Regression > Non-Linear Models

This utility performs computations that fit a non-linear mathematical regression model to two variables. To use this utility, you must provide the values for an independent/predictor variables and a dependent/response in separate columns.

The following types of non-linear models are available:

• Quadratic: y = ax² + bx + c
• Cubic: y = ax³ + bx² + cx + d
• Polynomial: y = axⁿ + bx^n-1 + cx^n-2 + ..., where n is a positive integer
• Logarithmic: y = a ln(x) + b
• Power: y = ax^b
• Fixed Power: y = axⁿ, where n is a real number
• Exponential: y = ab^x

where y is the independent variable with N sample values and x are the dependent variable with also N sample values. The regression coefficients are determined by least squares estimation.

For variation, the following values are computed ( is the mean of the y variable values, is the y value predicted by the regression equation, n is the sample size, and k is the number of regression coefficients):

• Explained variation:
• Unexplained variation:
• Total variation:
• Coefficient of determination:
• Standard error of estimate:
• Test statistics F:
• p-Value corresponding to the test statistics

Dialog Inputs

• Select x (independent/predictor variable): Select the column containing the x values
• Select y (dependent/response variable): Select the column containing the y values (the number of x and y values must be the same)
• Type of Model: Select desired type.