High School Student Grade Prediction

Which variables are important

Team Hitmontop
Jiaqi Liu (jl4424), Miles Ma (hm387), Yihong He (yh827), Linlin Li (ll966)

2023-12-04

Topic and Motivation

Analyze and predict student performances within high schools.

Investigate the connections between students’ grade and other variables present within the dataset.

Develop predictive models illuminating the factors influencing student performance.

Introduce the data

Datasets obtained from the UC Irvine machine learning repository, containing information about student achievements in secondary education from two Portuguese schools, covering the subjects of Math and Portuguese.

Merged two datasets
1044 rows and 35 columns

Data source: https://archive.ics.uci.edu/dataset/320/student+performance

(Ineffective) Visualizations

Highlights from EDA

ML model tried: null (3.09), random forest (1.99), svm (2.07), lasso (1.43, chosen)

Conclusions + future work

According to machine learning results, extra educational support and family educational support are the most important factors for students’ grades. Other important factors include mother education, past extra classes, etc.
Visualize how each individual’s grade is estimated
Introduce a shiny web to play and test our results from machine learning