Machine Learning Projects

• Extreme Learning Machine & Multilayer Perceptron

  Exploring the classification of handwritten digits from the EMNIST dataset using Extreme Learning Machines (ELM) and Multi-Layer Perceptrons (MLP). The project includes preprocessing and loading the EMNIST data, building an ELM model and a simple MLP model to classify the digits, and evaluating and comparing the performance of both models. View Project

• Constructing Convolutional Neural Networks on EMNIST

  Classification of handwritten digits from the MNIST dataset and handwritten letters is explored using Convolutional Neural Networks (CNN) and a pre-trained VGG16 model. The project includes preprocessing and loading the MNIST and letters data, building a custom CNN model and utilizing a pre-trained VGG16 model through transfer learning for classification, and evaluating and comparing the performance of both models. The models are trained and tested on both the MNIST digits and the letters dataset, with accuracy and loss visualizations provided for performance comparison. View Project

• Recurrent Neural Networks on Stock Price Prediction

  Stock price prediction for AMP (AMP.AX) is explored using Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) networks, and Gated Recurrent Units (GRU). The project involves retrieving historical stock data, pre-processing the data, and building custom RNN, LSTM, and GRU models to predict future stock prices. The models are trained and tested on data spanning from 2000 to 2020, with performance evaluated using Mean Squared Error (MSE). Visualization of predicted stock prices versus actual values is provided for each model to assess their accuracy. View Project