built a linear model comprising a dependent continuous variable, 3 continuous predictors, 2 discrete (factors with two gradations each), as well as the interaction of one of these continuous and one discrete predictor. I want to build a series of a series of graphs, reflecting the connection of the dependent variable with predictors, as well as on one of them to visualize the interaction. The course was taught to depict the connection with the most important predictors, provided that all other predictors are unchanged, for example, their values are equal to average. Only how to make the factors unchanged? Averaging them, understandable, will not work equal to zero -also does not accept.
-
Answer # 1
-
Answer # 2
The question is quite common, therefore the answer is the same.
Simple and unambiguous way to assess the relationship of a single predictor with a target variable within a particular model does not exist. There are a number of approaches that are discussed in detail in the book. https://ema.drwhy.ai/(In particular, see Break-Down Plots and Partial-Dependence Profiles).
- r : Change ggplot legend options
- Signatures for columns in geom_histogram
- R. Add correct legend to ggplot plot
- r - about handling geom_vline in ggplot2
- i don't know how to format data for graph drawing with r and ggplot
- coastline plot in r
- r - i want to use a function instead of opts in ggplot2
- i want to specify the bubble color using variables in the data frame with ggplot2 of r
- cannot create line graph with ggplot2
The question is quite common, therefore the answer is the same.
Simple and unambiguous way to assess the relationship of a single predictor with a target variable within a particular model does not exist. There are a number of approaches that are discussed in detail in the book. https://ema.drwhy.ai/(In particular, see Break-Down Plots and Partial-Dependence Profiles).