connormcmanigal / Recipe-Review-Scoring-Algorithm-Regression

Leveraging sentiment analysis and data augmentation to recreate recipe scoring algorithm with sparse data. Used MLPs and Gradient Boosting Regressors to compare regression metrics such as RMSE and MSE between raw data and raw data in conjunction with augmented data.

Recipe Review Scoring Enhancement

Leveraging Sentiment Analysis and Data Augmentation to Recreate Scoring Algorithm with Limited Data

Connor McManigal, Peyton Politewicz, and Harold Ng

View the project here: https://connormcmanigal.github.io/Recipe-Review-Scoring-Algorithm-Regression/report.pdf

Overview:

This project's primary objective is to predict the best score of a recipe review, a metric utilized by social media websites to determine the order in which comments appear. We create new features with the existing data and leverage VADER and TextBlob sentiment analysis libraries to obtain polarity and subjectivity scores reviews, aiming to enhance performance metrics through the incorporation of additional data. We used MLP and Gradient Boosting Regressors as our algorithms, and we evaluated their performance using error metrics such as Mean Absolute Error and Mean Squared Error.

Data Citation:

• Textual Taste Buds: A Profound Exploration of Emotion Identification in Food Recipes through BERT and AttBiRNN Models
  • Authors: Amir Ali, Stanisław Matuszewski, Jacek Czupyt, Usman Ahmad
  • Published: 2023
  • Journal: International Journal of Novel Research and Development
  • Volume: 8
  • ISSN: 2456-4184