Training a recommender system on a standalone dataset
In the Training a recommender system on a small curated dataset and Training a recommender system on a large curated dataset recipes, we used two curated datasets adapted from the MovieLens dataset that contains user IDs, movie IDs, and user ratings for these movies. Using these two curated datasets, we created recommender systems that predicted what rating a user would give to a particular movie.
In this section, we will explore a dataset that is not a part of fastai's set of curated datasets. The dataset we will use in this section is the Amazon product dataset (https://www.kaggle.com/saurav9786/amazon-product-reviews). This dataset contains user ratings for a large range of products available on Amazon.
In this section, we will be working with a subset of the dataset, the subset related to electronic goods. In the recipe in this section, we will ingest this standalone dataset and then use it to train a recommender...
