Upload numerical data → generate new predictions → store predictions → update clients.

Training

• Upload training and testing data to S3 [s3://cst3130-cw2-machine-learning]
• Use DeepAR algorithm and train model.

Used two data sets for training

Train - Data with approximately 30% of the values removed [synthetic_data_test.json]

Test: Complete set of data [synthetic_data.json]

Results: [s3://cst3130-cw2-machine-learning/results]

Prediction

• Deployed model to endpoint. Name: SyntheticEndpoint
  URL: https://runtime.sagemaker.us-east-1.amazonaws.com/endpoints/SyntheticEndpoint/invocations
• Query endpoint for predictions on the frontend