The Kaggle "What's Cooking" dataset provides a unique looks at different types of cuisines and the ingredients associated with them. My project trains a Support Vector Machine using Scikit-Learn to classify these cuisines based on their ingredients. The end goal of this project on the machine learning side is to help people find new recipes they might like based on their existing tastes. Given the finite accuracy of the algorithm, as determined during cross-validation, there are cuisines that are commonly confused with each other, which will be exploited to suggest new recipes to people. A goal more broadly is to provide a quick way to view the nutritional information for a given recipe.
christopher-w-murphy / whatscookingsvm Goto Github PK
View Code? Open in Web Editor NEWLicense: GNU General Public License v3.0