Applying transformations to the data

Data transformation is a crucial data normalization procedure that must be done before we feed the data to a neural network. We need to transform non-numeric features to numeric values and handle missing values. In this recipe, we will perform schema transformation, and create dataset iterators after transformation.

How to do it...

1. Add features and labels into the schema:

Schema.Builder schemaBuilder = new Schema.Builder();
 schemaBuilder.addColumnString("RowNumber")
 schemaBuilder.addColumnInteger("CustomerId")
 schemaBuilder.addColumnString("Surname")
 schemaBuilder.addColumnInteger("CreditScore");

2. Identify and add categorical features to the schema...