#
5. Make Predictions About New Data
attachment
Students can submit work

## Transcript

In previous videos, you used the training data set to develop a model that predicts how much someone will like an item.

Now you will use the test data set to predict the rating someone will give the first item listed in your Google Forms survey.

To do this, use another array formula.

Begin by inserting a new column to record the predicted rating.

Add a heading.

Next, use an array formula to add the “average” differences you calculated in the previous video to the “actual” rating the user gave items two through five.

In the first row of your test data, type “equals array.”

Select “array formula” from the menu.

Then, type “average” and select this formula from the menu.

Choose the average differences you calculated for the training data.

Then, type a plus sign ... And select the actual ratings the user gave for items two through five.

Type two closed parentheses to complete the average formula and the array formula.

Your formula should look like the one onscreen.

Press enter.

Terrific!

Your model returned a result!

If your model did not return a result, or if you got an error message, select the cell and check the formula.

Make sure the parentheses are in the right place and that you added the calculated differences...

To the responses...

And used the array formula to calculate these across the different rows.

The next step is copying the model to the other users in the test data set.

But first, add absolute cell references to the calculated differences.

These are the cells you do not want to change.

Now, drag the cell handle to copy the formula.

Check your work.

The next formula still uses the calculated differences, but changes based on the row number.

Great work!

You developed a model that can help you predict ratings from the test data.

Now, it’s your turn: Insert a column for the predicted rating.

Add an array formula that uses the predicted ratings from all of the test data users.

Calculate the predicted rating for the first user in the training data from all of the test data users.

And copy the model to the other users in the test data set.

- Insert a column for the predicted rating.
- Add an array formula that uses the predicted ratings from all of the test data users.
- Calculate the predicted rating for the first user in the training data from all of the test data users.
- Copy the model to the other users in the test data set.

*attachment*

This project will be shared with your teachers

Students can submit their work on this page. View their submitted work on the student progress page of My Classes.