All language subtitles for 5. Split the dataset

Hi, everyone, and welcome in this new video.

In this video, we're going to create our different sent to do our analysis.

First, we need to define how much data we want in the sense and amateur data we want in the test because

our algorithm needs it.

The train sets to calibrate each coefficient and a test set to test the performance of this algorithm

on UNO data.

So two of the variable splits we need to define the threshold to separate the train test and the test.

Usually, we take that officials at 80 percent.

It means that 80 percent of the data all used to train the algorithm and 20 percent of the data are

used to test the performance of this argument.

So to do it, we take the integral parts of.

The length of or that a friend multiplied by 90 percent.

Then we need to define that which are used as features.

So the variable that we are going to use to predict all targets.

So this dataframe is called X and we need to define X for the train and X for the test.

So.

For the train and the test for the X, we want to take the semi, the moving relativity and the RSI.

And then we have the property you look, we split.

The data.

And we do exactly the same thing for the white train, which contain the targets for the train sets.

Which is.

The returns of the asset and also we need to split taking only the train data.

Then we need to do exactly the same thing for the extremists and the widest.

The only difference is that now we have taken the data from the beginning to the split.

And for the tests, it's we need to take the data from the street to the end of the dataset.

And let me just show you some of our.

To be sure that you are really understand what we have done here.

So, for example, extreme contain all the necessary data to predict

wine train, which is the returns of the assets, and we have to do exactly the same thing for the tests.

And as we can see, we have exactly the same data, but not in the same time because we have this database,

we are going to train our algorithm.

And with this database, we're going to test or trading strategy following our linear regression prediction.