/ML-Tasks

Primary LanguageJupyter Notebook

ML-Tasks

this is a repository dedicated for the tasks required for the machine learning course at SBME department, Cairo university. The required tasks usually ask us to use sklearn models on different datasets and comment on the findings, and build a machine learning classificiation or regression model from scratch.

mnist data (task-1)

in this task I wrote a jupyter notebook that does binary and multi-class classification using an sklearn model on the mnist data. the confusion matrix is computed and the results are discussed in the notebook.

Building a linear regression model (task-2)

in this task, I implemented multivariate linear regression using gradient descent. the model has the following functions: • fit: method to train the multivariate linear regression model • predict: method to use the trained linear regression model for prediction • Compute Cost: compute the value of the objective function • Gradient Descent: optimizes the parameter values via gradient descent • Evaluate Performance: calculate the accuracy of the prediction for the test data.

it was tested on univariate and multivariate data, results are shown in the notebook.

building a CART Tree model (task-3)

in this task, I Implement the CART decision tree learning algorithm using both gini impurity and entropy.

building SVM model

in this task, I Implement the SVM learning algorithm with a linear kernel using gradient descent. the model has the following functions : 1- Fit: method to train the SVM model 2- predict: method to use the trained model to generate predicted labels.