iCurveFit: Interactive Curve Fitting

In mbojan/mbstats: Michal's Beautiful Selection of Totally Awesome and Trustworthy Statistical functions

Description

Interactively enter a set of points and then fit a curve of specified form to get a coefficients.

Usage

iCurveFit(
  formula = y ~ x,
  n = 5,
  model = "lm",
  prn = 100,
  xlim = c(0, 10),
  ylim = c(0, 10),
  ...
)

Arguments

formula	formula for the model to be fitted to the entered data
n	numeric, number of data points to be entered
model	character, name of the model-fitting function to be used for the entered data
prn	numeric, number of points used to evaluate the fitted model, important only for plotting the fitted curve
xlim	numeric of length 2, limits for the horizontal axis
ylim	numeric of length 2, limits for the vertical axis
...	other arguments passed to the model-fitting function

Details

Frequently there is a need to quickly come up with a particular form of a function. This function enables to quickly fit a curve of certain kind to a, usually small, number of data points.

The function starts with creating an empty plotting region ready for input of n data points. Every data input (through locator) results in adding a point to the plot with a subsequent number. After the data are inputed the function uses model-fitting function specified with model to fit the curve specified with formula with possibly additional arguments passed through ....

The formula argument has a usual form as in almost all model-fitting functions in . It can contain only variables y and x. Furthermore, direct transformations of y are not supported. All transformations of x should be enclosed in I() function.

The function was designed to be used with lm or glm as model-fitting functions. However, in principle, it should be possible to use any other function that provides similar interface, i.e. formula and data arguments, and the corresponding predict method.

Once the model is fitted the function finds its predicted values for a sequence of values defined by xlim of length prn. Then the curve itself is added to the plot. Finally the model object is returned.

Value

A model object created for the supplied data with the model-fitting function specified with model.

See Also

See lowess and spline for other curve fitting functions, lm and glm for help with th model-fitting functions supported. Also see predict and I

Examples

## Not run: 
# fitting via 'lm'
iCurveFit(formula= y ~ x ) # linear
iCurveFit(formula= y ~ I(x^2) + x ) # parabolic
iCurveFit(formula= y ~ I(x^3) + I(x^2) + x, n=7) # polynomial of degree 3
iCurveFit(formula= y ~ I(sin(x)), n=7) # sinusoidal
iCurveFit(formula= y ~ I(log(x)), n=7) # log
iCurveFit(formula= y ~ I(1/x), n=7) # hyperbolic

## End(Not run)

mbojan/mbstats documentation built on Dec. 21, 2021, 3:56 p.m.