This project was developed as part of the Machine Learning Program at Data Science Academy. The goal was to build a simple binary classification model that predicts whether a student will pass or fail based on three input variables.
The main objective is not model complexity, but understanding how to deploy a machine learning model to AWS. For that reason, the dataset and modeling approach were intentionally kept simple.
- Scikit-learn
- Pandas
- NumPy
- Pickle
- AWS (focused on EC2)
In this step, basic preprocessing was performed using StandardScaler. After scaling the data, three different classification models were trained and evaluated.
The selected model for deployment was Logistic Regression, chosen based on its performance and simplicity.
In the app.py file, a simple backend was built using Flask.
This application loads the trained model and serves predictions through an HTML template, allowing user interaction via a basic front-end interface.
The project was deployed to an AWS EC2 instance. This involved configuring the virtual machine, installing dependencies, and transferring the application files to the server.
The application was executed using a WSGI server inside the EC2 instance.
After configuration, the system became accessible through the public IP address of the instance, allowing users to interact with the deployed model via the browser.
This is a simple project focused on understanding the fundamentals of deploying a machine learning model on AWS.
The emphasis is not on building a complex model, but on learning how to structure, deploy, and serve machine learning applications in a cloud environment for future, more advanced projects.
For privacy reasons, steps that explicitly show the AWS instance configuration and environment details were intentionally omitted.