Sentiment Analysis instead of Positive Negative Neutral, I want to sentiment by category of products

Question

Im doing a categorical product sentiment analysis. But I dont quite know what keyword to search for and what methods should I use. Im using this dataset for this https://huggingface.co/datasets/viewer/?dataset=amazon_reviews_multi&config=en Im doint a neural search project and my model would be the Review text and product category.

Answer 1

I think the keyword you are looking for is 'multi-task classification'. Sometimes this is called 'joint learning', where a model learns how to solve two or more tasks simultaneously.

Task 1: Classify sentiment. Task 2: Classify product category.

It is possible to train one neural network model to learn how to classify both sentiment and category at the same time. You'll end up having an output layer with two groups of nodes, one for the first task, and one for the other. You need to define loss criteria for both groups, add them together, and use the combined loss to train the network.