This repository contains the code and documentation for a deep learning lab that focuses on regression and multi-class classification tasks. The project involves predicting stock prices and performing multi-class classification on a manufacturing process dataset.
During this project, we addressed two main tasks: regression and multi-class classification. In the regression task, the goal was to predict stock prices using a deep neural network implemented with PyTorch. The multi-class classification task involved a comprehensive analysis of a manufacturing process dataset, including data preprocessing, EDA, model development, hyperparameter tuning, and evaluation.
In the first part, we implemented a deep neural network to predict stock prices. The notebook DL_lab1_part1.ipynb details the data preprocessing, model development, hyperparameter tuning, and visualizations.
The second part involves multi-class classification on a manufacturing process dataset. The notebook DL_lab1_part2.ipynb covers data preprocessing, exploratory data analysis (EDA), model development, hyperparameter tuning , and performance evaluation.
Visualizations of loss and accuracy over epochs are available in the respective notebooks. The project also calculates metrics such as accuracy, sensitivity, and F1 score for both training and test datasets.
Throughout the project, valuable insights and skills were gained, including proficiency in PyTorch, effective model development, and rigorous evaluation techniques.