Try other models, including nonlinear models, supervised AI models, decision tree, clustering methods etc.

If you are interested, send me the data to see.

[email protected]

I think you have a lot of Outlier data

Delshad Shaker Ismael Botani, it sounds as if you are asking for advice about how to "torture the data" until it confesses.

• https://quoteinvestigator.com/2021/01/18/confess/

Why is it not possible that the linear combination of your 4 explanatory variables that minimizes the sum of squared residuals simply does not account for much variance in Y?

Here are some additional questions that come to mind:

HTH.

PS- In the statistical dialect I speak, your model is a multivariable model, not a multivariate model. The latter term describes a model with two or more dependent variables. You have only one dependent variable, but 4 explanatory variables. Thus, it is a multivariable, univariate model.

Article Multivariate or Multivariable Regression?

Dear Bruce Weaver

As we know R2 = 1- (SSE/SST). So when R2 is very low, it looks like SSE is almost equal to SST. that's mean

Sum (yi - mean(y))2 almost equal Sum (yi - Estimated(yi))2 →

(mean(y) - Estimated(yi))2 almost equal 0. →

mean(y) almost equal Estimated(yi).

One of the most common causes of this to happen is when there is Outliers in the data.