This repository contains the analysis made for the Machine Learning Competition for the data analytics bootcamp, at Ironhack Barcelona (March,2021).
The challenge consists in predictig the quality of cookies. We received a dataset with 16 columns of cookie measures (like sugar to flour ratio, weight, diameter,..) and the quality of each cookie. The goal was to use supervised machine learning techniques to be able to predict the quality of another set of cookies. The metrics used to evaluate the quality of the predictions was RMSE.
The general workflow can be splitted into 5 main steps:
First step: understanding the meaning of every column Second step: exploring data and looking at values: do they all make sense? Third step: cleaning data (dealing with missing values, checking correlations, outliers...) Fourth step: performing machine learning strategies and checking for the quality of each model Fifth step: trying different parameters and techniques to get to the optimal prediction result
In this repository you'll find:
- Notebook: The Jupyter Notebooks with the ML models: The one labeled "final" has the ultimate predictions.
- Data: datasets used for the predictions
- Figures: graphs used to do some modeling