Man is What He Eats: A Research on Hinglish Sentiments of YouTube Cookery Channels Using Deep learning

Abstract

Our study focuses on the sentiment analysis of Hinglish comments by multi-label text classification on cookery channels of YouTube using Deep learning. Multi-layer perceptron (MLP) with different parameters was implemented in our study to investigate the various sentiments in the comments. We have modelled and evaluated MLP by varying the number of neurons, layers, optimizers, activation functions with the various feature engineering methods such as tf-idf, count vectorizer, pre-trained embeddings and customizedembeddings. These experiments were conducted on two datasets they are Kabita’s Kitchen and NishaMadhulika’s dataset. From the investigation, we concludedKabita’s Kitchen dataset has the highest accuracy 98.53% and NishaMadulika’shas 98.48% accuracy in MLP.This outcome of the experiment was evaluated based on careful analysis on tests conducted during our study