How to plot a partial regression plot for logistic regression (using fitglm)?

Question

Prashanti Ganesh 2022-7-18

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1762885-how-to-plot-a-partial-regression-plot-for-logistic-regression-using-fitglm

回答： Aman Banthia 2023-9-13

Dear All,

I am using a logistic regression model with a few predictors and a binary response model. I would like to a plot a partial regression plot for such a model. However, it seems that there is a function to plot partial regression only for linear regression. Is there an alternative of that kind for logistic regression?

Thank you.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Aman Banthia 2023-9-13

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1762885-how-to-plot-a-partial-regression-plot-for-logistic-regression-using-fitglm#answer_1308996

在 MATLAB Online 中打开

Hi Prashanti,

I understand that you are trying to plot a partial regression plot of a for logical regression model using ‘fitgml’ function.

Plotting a partial regression plot for logistic regression using the `fitglm` function in MATLAB can be a bit more involved compared to linear regression. Here's a step-by-step approach to create a partial regression plot for a single predictor in a logistic regression model:

1. Fit the logistic regression model: Start by fitting your logistic regression model using the `fitglm` function. Make sure to specify the appropriate family and link functions for logistic regression, such as 'binomial' for the family and 'logit' for the link function.

% Fit the logistic regression model
mdl = fitglm(X, y, 'Distribution', 'binomial', 'Link', 'logit');

Replace `X` with your predictor variables and `y` with the binary response variable.

2. Create a range of values for the predictor: Choose a range of values for the predictor variable for which you want to create the partial regression plot. You can use the `linspace` function to generate a sequence of evenly spaced values within a specified range.

% Create a range of values for the predictor
x_range = linspace(min(X), max(X), 100);

Adjust the number of values (100 in this example) based on the desired resolution of the plot.

3. Generate predicted probabilities: For each value in the range of the predictor variable, generate the predicted probabilities using the `predict` function on the logistic regression model object.

% Generate predicted probabilities for each value in the range
predicted_probs = predict(mdl, table(x_range', 'VariableNames', {'Predictor'}));

Replace `'Predictor'` with the name of your predictor variable.

4. Plot the partial regression plot: Finally, plot the partial regression plot by plotting the predicted probabilities against the range of the predictor variable.

% Plot the partial regression plot
plot(x_range, predicted_probs, 'LineWidth', 2);
xlabel('Predictor');
ylabel('Predicted Probability');
title('Partial Regression Plot');