Creating multiclass batch predictions

- [Instructor] So in this video,…we're going to head back to the Machine Learning dashboard…and create a new datasource.…So I select the Amazon Machine Learning.…I select the dashboard.…I select create new, and datasource.…I provide the location of my test.csv files,…so that's ml-forest-type…and you see it's test-small.csv.…I give that datasource a name, so Forest Test,…select verify, select continue,…and we can confirm that the first line…contains the column names.…

Now if you remember, from the training set,…anything that had the word soil in it,…was of a type binary.…So let's select soil.…Select all of the entries with soil.…Let's view 100 items per page.…And let's change these all to type binary.…In the same way,…anything of type wilderness was also of type binary,…so I'm going to select all of them,…I'm going to change them to type binary,…and I'm going to select continue.…

So now, this test data set is in the same format…as our training data set,…and they should have the same schema.…So we shouldn't have the problems that we did…