Stats help - Linear Model fitting in R

Click For Summary
SUMMARY

The discussion focuses on issues with linear model fitting in R, specifically regarding a residual vs fitted values plot that exhibits a diamond-shaped variance pattern. The user attempted various transformations, including raising the response variable to a power less than one and applying log transformations, but these did not resolve the issue. Expert advice suggests that the linear trend in the residual plot indicates that a response transformation is unlikely to be effective. The conversation emphasizes the importance of understanding the number of explanatory variables and the modeling approach used.

PREREQUISITES
  • Familiarity with R programming and its statistical modeling capabilities
  • Understanding of linear regression concepts and residual analysis
  • Knowledge of data transformation techniques, including power and log transformations
  • Experience with interpreting residual plots and variance patterns
NEXT STEPS
  • Explore R's built-in functions for residual diagnostics, such as `plot()` and `residuals()`
  • Learn about advanced transformation techniques in R, including Box-Cox transformations
  • Research the implications of multicollinearity in linear regression models
  • Investigate alternative modeling approaches, such as generalized additive models (GAMs) in R
USEFUL FOR

This discussion is beneficial for data analysts, statisticians, and researchers working with linear regression models in R, particularly those facing challenges with residual analysis and variance stabilization.

aerot89
Messages
4
Reaction score
0
I have this residual vs fitted values plot in R, and it should show a scatter of points with constant variance and mean 0. The mean 0 part seems to hold, but there is quite a clear diamond shape for the variance...

I have tried raising the response variable to a power <1 and a log transformation, along with transforming all of the parameter types but all to no avail.

Any help would be greatly appreciated. Anyone have any tips as how to try and improve this...

Thanks

QpY3Z.png
 
Physics news on Phys.org
How many explanatory variables do you have? What are you trying to model? Are you restricting your intercept to a value of zero? From the clear linear trend in your residual plot, a response transformation will not help your situation. What is the R code that you are using?
 

Similar threads

Undergrad Linear Models vs Nonlinear Models
  • · Replies 22 ·
Replies
22
Views
4K
Undergrad Linear Model with independent categorical variable
  • · Replies 5 ·
Replies
5
Views
2K
Undergrad What is the expected value of Cov(x,y)2 in an independent X and Y scenario?
  • · Replies 17 ·
Replies
17
Views
3K
Undergrad Fitting two models in sequence with parameters in common
  • · Replies 12 ·
Replies
12
Views
3K
Undergrad What are the commonly used estimators in regression models?
  • · Replies 23 ·
Replies
23
Views
4K
Undergrad Nonlinear Least Squares or OLS for Nonlinear Models?
  • · Replies 7 ·
Replies
7
Views
3K
Undergrad Which method should I use for a linear fit of data with error on both z and y?
  • · Replies 8 ·
Replies
8
Views
2K
Undergrad About chi-squared and r-squared test for fitting data
  • · Replies 5 ·
Replies
5
Views
9K
Undergrad Linear fit on the difference of data points
  • · Replies 26 ·
Replies
26
Views
3K
Undergrad Best way to quantify how well a model fits data
  • · Replies 20 ·
Replies
20
Views
3K