Nima0803 / F21DL---Portfolio

F20DL Group 6 Portfolio

This file explains the directories in this repository.

Overview:

The Lab Folder consists of 7 Labs:

1. Lab 1: Dataset Exploration on 3 datasets (Spotify Top Hit Playlist (2010-2022), Water Quality and Potability and MNIST).
2. Lab 2: Data Exploration & Visualization on previous 3 datasets by generating graphs.
3. Lab 3: Selected Spotify dataset and applied Feature Reduction and PCA.
4. Lab 4: Explored an image dataset called JMuBEN2 (Arabica coffee plantation leaves), applied pre-processing and normalization, trained a CNN model with 50 epochs and an SGD Classifier, and performed Evaluation Metrics on both models.
5. Lab 5: Instead of using CNN and SGD Classifier on the previous dataset, different types of Naïve Bayes Classifiers were trained on the dataset, and Evaluation Metrics were recorded.
6. Lab 7: Applied different clustering algorithms to detect the optimal number of clusters.
7. Lab 11: Exploring effects of different CNN architectures on the JMuBEN2 dataset.

The CW Folder consists of 4 Parts:

All parts involve exploring different techniques on a subsect of the German Street Sign Recognition Dataset.

1. Part 1: Data Visualization and Bayes Net
2. Part 2: Clustering
3. Part 3: Decision Trees
4. Part 4: MLP and CNN