This repository was created to share my solutions for the first 02 assignments done on the Machine Learning and Pattern Recognition class in my master's degree. All the problems presenteded on this class had some requirements and restrictions regarding strategies and techiniques that imposed sometimes a pre-defined methodology to be followed. All of those points are discussed on the reports.
For each task there is a jupyter notebook containing the solution developed and a writen report explaining step by step of how the results were achivied as well as the dificulties/limitations found on each problem. You can find those reports on the following links:
-
The goal of this practice is to explore possible solutions for a regression problem using different approaches based only on a simple Linear Regression. The dataset was published by Mashable and preprocessed before given to the class through natural language processing techniques. The final goal is to predict the number of shares in social networks of a given set of articles based on a set of 58 different features.
-
Assignment 02: Classification;
The goal of this practice is to explore possible solutions for a Kaggle competition. The competition is an image classification problem aiming to predict the camera models in which the photos were taken. There are a total of 10 types of cameras models and the training dataset is composed by 275 pictures for each class and a test set of 2640 images not labeled.