Linear Regression Descriptive Model

[akaditya6896]

Hi, I want to build linear regression descriptive model which would show me the estimates, p values, std. error, R square…etc and also the diagnostics of the model (Residuals vs Fitted, qq plot, residual noramality… like we get in R). I have tried the linear regression model but in the output it gives only the variables estimates and the scatter plot of y vs independent x. Kindly help me on how to get the diagnostic plots in KNIME.

[lisovyi]

Hi @akaditya6896,

welcome to the KNIME forum community!

There is no single automatic tool prepared to generate all the information that you are interested in. There are two options:

1. You can calculate all the relevant statistics using KNIME nodes, For example, using corresponding Regression Predictor you can predict on a hold-out test set, using Numeric scorer you can calculate various statistics like R^2, MAE, etc and using math operations and plots you can visualise everything as various distributions. If you generate such a dashboard and make it general and configurable, you are invited to share it on the KNIME Hub, such that other users can benefit from your work.
2. You can use R (or python) integration. This way input KNIME table will be transformed into an R dataframe and you can do various manipulations on it. You can even generate view using R view, for example using

https://hub.knime.com/knime/extensions/org.knime.features.r/latest/org.knime.r.RViewNodeFactory

Best regards,
Mischa