Deploy a Sentiment Analysis Model

Udacity Machine Learning Engineer Nanodegree - Project 1

Project Introduction

This project is a small Web App, based on Amazon SageMaker, which can be used to easily and quickly predict whether a review is positive or negative.

Example

Here is a review example for the film: The Personal History of David Copperfield

These review was posted on 26 January 2020 by Mark Kermode and was rated by him with 5 stars, therefore the review is to be rated POSITIVE.

With the Sentiment Analysis Web App, the review text can be analyzed to determine whether it is positive or negative (Assumes that the corresponding model endpoint is enabled in Amazon SageMaker!). The original review text can be cut and pasted into the review input field of the Web App. When you click on the Submit button, the system displays a prediction as to whether the review is to be considered positive or negative:

Model Building and Deployment

The Jupyter Notebook SageMaker Project explains how the model was created, tested and deployed using Amazon SageMaker. The result is also shown in the report.html file.