Feature extraction of 'real' and 'fake' images and implementation of the best classification method (using various machine learning models such as Random Forest, SVM, and Logistic Regression) to identify fake images.
Conducting one-vs-all classification on the 'penguins.csv' dataset using both the Naive Bayes classifier implemented from scratch and the one provided by Scikit-Learn's built-in functions.
Performing a Decision Tree classifier on the 'Diabetes.csv' dataset to distinguish between diabetes and non-diabetes cases.
Implementation of EM algorithm, and performing GMM estimation on an image dataset containing images of Manchester United and Chelsea football clubs to classify them.
Performing regression of several degrees on generated data points affected by white and Poisson noise. Additionally, studying its MSE loss and bias-variance trade-off.
Implementing logistic regression with L2 Regularization from scratch to classify two circular datasets.
Performing an MLP model on the ECG signals dataset.
Performing a CNN model on the CIFAR-10 image dataset.
Performing PCA on the "emotion_detection" dataset contains 213 images with 6 labels: Happy, Fear, Angry, Disgust, Surprise, and Sad.
Performing non-parametric Parzen Window estimation on the 'ted_main.csv' dataset, both from scratch and using built-in functions.
Performing Support Vector Classification (SVC) on the iris dataset.
Implementation and analysis of K-Nearest Neighbors with varying values of K, including a study on the impact of metric learning using LMNN and LFDA algorithms.
Implementation of K-Means++ and K-Medians++ clustering algorithms, including an analysis of their performance across various values of K.