CertificateBackground.png

Certificate of Completion

ErdosHorizontal.png

THIS ACKNOWLEDGES THAT

HAS COMPLETED THE SPRING 2022 DATA SCIENCE BOOT CAMP

Kalven Bonin

clear.png

Roman Holowinsky, PhD

JUNE 08, 2022

DIRECTOR

DATE

TEAM

Gordon Ramsey

Ronak Desai, Kalven Bonin, Shidhesh Supekar

clear.png

Is it possible to classify a recipe’s cuisine type just from a list of
ingredients? Our project seeks to answer this question and
does so using some basic tools of Natural Language Processing.
We take a dataset from Kaggle.com that has a list of ~40,000
recipes with a cuisine type classification. One method we
employ is called a Bag of Words (BoW) model which take all
words found in the ingredients list and builds a classifier based
on the occurrences of those words in the training set. The other
method is the Term Frequency-Inverse Document Frequency
(TF-IDF) which considers the frequencies of individual words in
the training set. Both methods produced a testing accuracy of
greater than 60%, which is good considering that we
implemented the most naïve NLP models.

Screen Shot 2022-06-03 at 11.31.35 AM.png
github URL