/MLModelDeployment

An auto-scaling ML application to predict medical expenditures hosted on AWS.

Primary LanguageJupyter Notebook

Project Description

This project is a containerized auto-scaling ML application to predict medical expenditures hosted on AWS.

You can view a project overview video here.

A more detailed written description of the project can be found here.

User Interface

Index Page

Index

Prediction Page

Prediction

Sample Web API Call

Here is a sample call to the API with the following parameters:

Age: 23

BMI: 25

Number of Children: 3

Sex: Male

Smoking Status: Yes

Region: Southwest

Host IP Address: 54.224.27.206

curl -v -H "Content-Type:application/json" -X POST -d "{\"age\":23, \"bmi\":25, \"children\":3, \"female\":0, \"male\":1, \"no\":1, \"yes\":0, \"northeast\":1, \"northwest\":0, \"southeast\":0, \"southwest\":0}" 54.224.27.206/:8080/predict

Here is the response:

{"prediction":[3284.857731772743]}

Architecture

Architecture

Team Members

Derek Wales, Duke MIDS

Joe Krinke, Duke MIDS