A regression of 2 portuguese classes using bayesian statistics, Keras deep learning other ML models.
Jupyter Notebook 100.00%R 0.01%
studentoutcomepredictor's Introduction
StudentOutcomePredictor
Predicting students final grade with Keras Deep Learning and a comparison of Machine Learning Models
Understanding the data better with Bayesian Statistics
Overview
The dataset obtained from UCI Machine Learning Repository. original dataset
Two Portuguese high schools surveyed students in a Math and a Portuguese language classroom. Features such as demographic, social, home-life and past grades were collected from students directly and assumed to be accurate.
The original study Cortez and Silva, 2008 used machine learning models. This project takes the study a step further by using Keras Deep Learning.
Bayesian Statistics are used to create a deeper understanding for the impact of life-choices and life's circumstances on student performance. While it would be unethical for a school to use the ML models for admissions a look at the bayesian statistics from the collected data is ideal for educating families and students on how to improve their likelihood of success.