SciPy Recipes

Creating a sparse matrix

In order to understand sparse matrices, we will consider the following real-world scenario: recommending the next item that a supermarket customer is likely to buy, given a set of historical transactions.

In a typical supermarket, there can be millions of customers and thousands of items. Any given user would have bought only a few items among the thousands of items present in the supermarket.

We can represent all the transactions of a supermarket in such a way that all the customers are represented in rows and all the items are represented in columns. The cell values are 1 if the customer bought the item, and 0 otherwise.

In the preceding scenario, we will have a very high majority of zeros and very few ones. This ...