Linear-Regression-Model | simple Linear Regression Model in Machine Learning | Machine Learning library

My problem is for the mtcars data set in R, I need to create all possible additive linear regression models where I'm regressing on the mpg variable. The null model is easy, as there's

10 choose 0 ways to get the null model, and 10 choose 1 ways to create a SLR on mpg; 10 choose 2 ways to create a two variable regression on mpg; 10 choose 3 ways to create a SLR on mpg; etc.,

So in total, as this is equivalent to summing across the 10th row in Pascal's Triangle, the total models I need to consider comes out to be 1,024.

Now, the other tricky part is I need to somehow store each model in some separate object so that all the 2 variable models are grouped together, all the three variable models are grouped together, etc, on top of also storing all them together (though perhaps there's a more efficient way to do this). The reason for this is my task is to look at all of these models, take their AIC scores and their Mallow's Cp scores, store those in a data frame, and then sort those scores from lowest to highest and keep the top 10. On top of this, I need to also be able to store, see, and have access to/use the two best 1-variable models through the two best 10-variable models because I need to provide the r-squared values and adjusted r-squared values for these various models along with the error mean square value. I'm still pretty/relatively new to R/coding in general, but I provide my attempt below:

I am trying to run a regression in a loop with variables names changing at each loop. Similar to this setup here

At the end, I would like to save the fitted results in a list.

My code is the following:

I have a cummulative rainfall time series and I would like to detect the change points. Here's the data.

So I have a data set that has 188 rows and 65 columns relating to World development indicators and Birth statistics. I am trying to do a purposeful selection method to create a regression model. The first step of this is to look at all of the individual simple linear models.

my goal is to run regression models in R for for each of my variables against my response. I know I can run lm(x$v30 ~ x$v1) which would give the regression for one of the variables. however, i am hoping to be able to do this in one step and pull all of the p values into a table or write them to a CSV.

I was following this but this does not give the P-values in a nice manner:R loop for Regression

This might be a little bit of a silly question (and probably trivial question) but I am new to machine learning. This can probably be easily deduced from the code I came up with, and it's not an excuse for a poorly formulated question. If you find this question poorly formulated, please inform me so I can update it.

I trained a multiple linear regression model and I want to see how well it performs a given data set. So, I googled around and I found a nice article explaining me how to find out the "error" of the predicted values, from the true ones. A couple of options it gave me were:

I applied all of them and they gave me incredibly high values, so I don't know whether these are correct or how I should interpret them.

Output the article was receiving:

• 10.0
• 150.0
• 12.2474487139

Output that my model received:

• 7514.293659640891
• 83502864.03257468
• 9137.990152794797

As a quick reference, these are my true / predicted values

The 'TLDR' question: Am I measuring my error correctly using the above mentioned methods, and are these results implying that my model performs incredibly bad? (This didn't seem like it, when I compared the prediction with the true values)

Here you can have a look at the data set I am using.

The code I used to create the model and predict values ( I tried to remove the unneeded code )

I use Matlab's stepwiselm to find a fit to my training data. The result model has a "total p-value" of the f-statistics that Matlab shows it. Now I want to use this model on a test dataset and calculate its p-value to determine how significant the prediction is.

Matlab has 3 commands to evaluate the model on a new data (link). But none of these commands automatically report a f-statistic p-value. Using predict and the model from stepwiselm for instance, returns the predicted response. Now the question is how I can find the p-value from the predicted values of the test set and their true values.

Thanks.

Here I asked how to compute AIC in a linear model. If I replace LinearRegression() method with linear_model.OLS method to have AIC, then how can I compute slope and intercept for the OLS linear model?