Lydia Farag's Projects
Using python and Keras, I am building a classifier that will train on 80% of a breast cancer histology image dataset (IDC dataset). 10% is retained for data validation (ie., error checking our machine learning model). I defined a convolutional neural network that is trained on the images and then I derived a confusion matrix to asses the model's efficacy.
HTN 2022
Created a common interface for two security interfaces that may be used by a **hypothetical** agency. These key encryptions are called Ouroboros and Cerebus and they are used based on the security level of the message.
I used the free API from Quandl which provides data regarding the SNP500 stock market up until 2018. Using the sklearn library and matplotlib, I used the linear regression functionality to predict the next 30 days with a very strong confidence (~0.9) of the machine learning. While antiquated data, a fun and hypothetically very useful piece of software.
A very basic foray into machine learning with python using logistic regression.
Config files for my GitHub profile.
Currently *under construction*. Using the Genius API, BeautifulSoup, and Markov Chains probability model, I'm developing a web app that will scrape lyrics for a user-inputted artist and generate fake (but veryyyy familiar) song lyrics.
*A work in progress* Using HTML, CSS (with SASS), and Javascript, I am finally working on building my personal portfolio.
Team Best Project
Need to settle the score? Say no more!
I created a text-based RPG in Java that implements the Factory and Decorator Design Patterns.
I used Java to create some code that optimizes student scheduling software; it allows for adding courses, verifying if the course can coincide with previously added courses without any time conflict, and tries to minimize the number of days with classes.
A silly experiment in using Bootstrap 5, CSS, and HTML5 to create a dating app knock off!