Dear experts, I am trying to obtain a function from a model, so that I could further manipulate it, plot it, etc. I can get model estimates and manually construct a function, but this gets tedious when trying out different functions to fit the data. There must be a better way of doing it, no? x <- c(1:10) y <- c(1:10) fit <- lm(y ~ x) f <- function(x){fit$coef[1] + fit$coef[2]*x} # Manually constructing function # Would be nice to do something like this: # f<-getFunction(fit) plot(f, 0, 10) Thanks, Alex Stolpovsky ----------------------------------------- This transmission may contain information that is privileged, confidential, legally privileged, and/or exempt from disclosure under applicable law. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution, or use of the information contained herein (including any reliance thereon) is STRICTLY PROHIBITED. Although this transmission and any attachments are believed to be free of any virus or other defect that might affect any computer system into which it is received and opened, it is the responsibility of the recipient to ensure that it is virus free and no responsibility is accepted by JPMorgan Chase & Co., its subsidiaries and affiliates, as applicable, for any loss or damage arising in any way from its use. If you received this transmission in error, please immediately contact the sender and destroy the material in its entirety, whether in electronic or hard copy format. Thank you. [[alternative HTML version deleted]]
On Jul 9, 2009, at 6:18 PM, Alexander V Stolpovsky wrote:> Dear experts, > > I am trying to obtain a function from a model, so that I could > further manipulate it, plot it, etc. I can get model estimates and > manually construct a function, but this gets tedious when trying out > different functions to fit the data. There must be a better way of > doing it, no? > > x <- c(1:10) > y <- c(1:10) > > fit <- lm(y ~ x) > f <- function(x){fit$coef[1] + fit$coef[2]*x} # Manually > constructing function > # Would be nice to do > something like this: > # f<-getFunction(fit) > plot(f, 0, 10) > > Thanks, > Alex StolpovskyIf you want to get the model fitted y values, just use: fitted(fit) As an aside, to get the model coefficients, there is also: coef(fit) and coef(summary(fit)) which like fitted(), is one of several 'extractor' functions that can be used on model object to get specific components. If you want to generate model predicted y values based upon 'new' x values, use: newdata <- data.frame(x = YourNewValues) predict(fit, newdata = newdata) Note that the 'newdata' data frame must contain columns with the SAME names as the independent variables in your original model. See ?predict.lm for more information, which can also generate various intervals, etc. If you want to plot your original data in a scatterplot and then add the model fitted line, use: plot(x, y) abline(fit) See ?abline for more information there. BTW, much of this is covered in An Introduction to R, which is included in your R installation and on the main R web site under Manuals link. HTH, Marc Schwartz
Thank you much for the useful pointers. Turns out, I can get what I want with predict(fit, newdata = newdata). But for the record, there is no way to get the fit function, from lm object, is there? Alex ----------------------------------------- This transmission may contain information that is privileged, confidential, legally privileged, and/or exempt from disclosure under applicable law. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution, or use of the information contained herein (including any reliance thereon) is STRICTLY PROHIBITED. Although this transmission and any attachments are believed to be free of any virus or other defect that might affect any computer system into which it is received and opened, it is the responsibility of the recipient to ensure that it is virus free and no responsibility is accepted by JPMorgan Chase & Co., its subsidiaries and affiliates, as applicable, for any loss or damage arising in any way from its use. If you received this transmission in error, please immediately contact the sender and destroy the material in its entirety, whether in electronic or hard copy format. Thank you.
On Jul 9, 2009, at 8:20 PM, Alexander V Stolpovsky wrote:> Thank you much for the useful pointers. Turns out, I can get what I > want with predict(fit, newdata = newdata). But for the record, there > is no way to get the fit function, from lm object, is there? > AlexAlex, There is no function returned within the model object. You can review the structure of the returned model object by using: str(fit) HTH, Marc Schwartz