Final Project Machine Learning
Abstract
The concept of beauty in both physical items and other people has long been a realm relegated to human judgment, a slave to human taste, and constrained by culture. However, regardless of its ephemeral and idealistic nature, beauty has real implications, as selling an unpleasantly appearing product will lose money, or in modern times, will make dating profiles less swiped on. The aim of this project is to introduce a ranking engine designed specifically for evaluating attractiveness in dating app photos. With the aim of assisting users in optimizing their profiles, this system applies recent gains in machine learning to create a ranking algorithm. The scale ranges from 1 to 5, offering straightforward feedback for users to enhance their profile presentation and increase their chances of successful matches. The project will leverage certain algorithms to refine and improve the accuracy of these recommendations continually.
By Michael Hakizumwami
DataSet:
https://www.kaggle.com/datasets/pranavchandane/scut-fbp5500-v2-facial-beauty-scores
To run,download the dataset and run all the cells in order, and ignore the ones that error out. Before running the 3rd cell in the new architecture section, run the last cell, and then run the rest of the cells.