The project consists form four tasks.
With the first task, each team needs to ensure the dataset is cleaned and the features are on the same scale.
The teams will apply an artificial neural network (ANN) in the second task. The teams have a completely free-to-choose number of hidden layers and the number of nodes on each hidden layer such that the number of hidden layers and nodes helps to satisfy the best performance to model. If the dataset will be implemented in an experiment that is not divided between training and testing sections, it recommends dividing the dataset as 70% for training and 30% for testing.
The third task will focus on applying a Convolution neuronal network (CNN). In this task, the Adam optimizer for optimizing and max pooling will be implemented as hyper-parameter coefficients. In addition, the teams have complete freedom to choose the size of the filter to implement. The experiment of the second and third tasks will be evaluated under the Accuracy, precision, recall, and F-score measurements.
In the last task, the teams must compare the results of the second and third tasks. In addition, the teams will explain the main reasons for performance differences between the second and third tasks experiments.
Out teams's given dataset : Hair loss dataset (https://www.kaggle.com/datasets/sithukaungset/hairlossdataset)
Verina Michel
Ola Mamdouh
Sandra Fawzy
Maria Anwar
Maria Mansour