Julia Fairbank CSCI 0451 Blog

Source code

https://github.com/CeceZiegler1/ML_Final_Proj/blob/main/LinearRegressionAnalytic.py

When using Recursive Feature Elimination, we need to select a few variables. First, we need to select what model we want to use. Since we are using multiple linear regression, we select linear regression. We also need to select how many features we want to use, and how many we want to remove during each recursion. These we chose arbitrarily for this example to be 10 and 1 respectively.

We also chose to implement and train a second model on our data. We implemented Random Forest Regression, since it is a very powerful model for making predictions with continous data. When we train this model on our whole data set, we get a much better validation score than with linear regression. However, just like with linear regression we can use feature reduction to reduce overfitting.

Predicting Bat Speed

CSCI 0451 Final Project: Create a machine learning model that predicts a players bat speed utilizing Recursive Feature Elimination, Linear Regression and Random Forest Regression models.
May 14, 2023
Cece Ziegler, David Byrne, Julia Fairbank, Sam Ehrsam

Auditing Allocative Bias

A blog post discussing the fairness of a prediction algorithm with respect to demographic characteristic bias.
May 8, 2023
Julia Fairbank

Learning from Timnit Gebru

A blog post reflecting on the work of Timnit Gebru.
Apr 18, 2023
Julia Fairbank

Linear Regression

Implementing least-squares linear regression and experimenting with LASSO regularization for overparameterized problems, CSCI 0451.
Mar 27, 2023
Julia Fairbank

Kernel Logistic Regression

Implementing kernel logistic regression, a method for using linear empirical risk minimization to learn nonlinear decision boundaries, CSCI 0451.
Mar 13, 2023
Julia Fairbank

Logistic Regression

Implementing simple gradient descent and stochastic gradient descent, comparing their performance for training logistic regression, CSCI 0451.
Mar 6, 2023
Julia Fairbank

Perceptron

Implementing the perceptron algorithm using numerical programming and demonstrate its use on synthetic data sets, CSCI 0451.
Feb 28, 2023
Julia Fairbank
No matching items