Effective Amazon Machine Learning

Creating an improved datasource

We need to do some manipulation on the new Titanic dataset before we upload it to S3 and create a new datasource in Amazon ML:

1. Open this new Titanic dataset in your favorite editor.
2. Select the first 1047 rows, and save them to a new CSV: ext_titanic_training.csv.
3. Select the next 263 rows and the header row, and save them to a file ext_titanic_heldout.csv.

We need to update our schema. Open the schema file titanic_training.csv.schema, and add the following lines to the JSON:

{  "attributeName" : "is_age_missing",  "attributeType" : "BINARY"  }, {  "attributeName" : "log_fare",  "attributeType" : "NUMERIC"  }, {  "attributeName" : "title",  "attributeType" : "CATEGORICAL"  }, {  "attributeName" : "deck", "attributeType" ...