Book recommendation system

In this practical section, we will build a recommendation system that will suggest different books to its online users, similar to the system we see on Amazon's website. Before we begin creating a recommendation system, we first need a dataset. Once we have one, we will then analyze that dataset as per our requirements and then apply our findings to recommend different items to our users.

Let's get started by setting up the dataset for our book recommendation system.

Dataset

At this stage, we should fully understand the power of data. If you don't have data, or have very sparse data, than it is almost impossible to build a recommendation system. Note that your data needn't be collected online – it can be the result of market research, user polling, and so on.

In this example, we will use the book rating dataset that was compiled by Cai-Nicolas Ziegler. This dataset contains the data of 270,000 books as rated by 90,000 users. The overall number of ratings it contains...