These programs are to demonstrate my abilities to solve classification problems given by Manchester Uni. with NN.
It has three distinct parts: Myopia & TREND problems. Links : Myopia - https://shorturl.at/ASTY5 TREND - https://zenodo.org/records/4521044 (taken only segmented files and copied to images/old & images/young, respectively)
Myopia problem (finished): ManchesterNN.py is a simple NN without any hidden layer based on data in Myopia_Data.xlsx ManchesterNN_opt.py provides an example of how some of the hyperparameters could be optimized
TREND (finished):
Here are two possible ways to work with the problem :
1. Using only files attributed to young people and their separation into healthy and obesity groups
2. Combination of young and old to find their differences.
Note! It is a small number of files, especially for older people. To resolve it, the file extensionDataset.py was added.
Additional. As a part of the work, training with Streamlit was add