Identifying the comparator pattern

The custom compare technique can be used to solve some specific types of problems. These are generally medium or hard problems where we need to find the aggregated value of a function over the top k items in a dataset, where each item is of a user-defined type. Most problems solved using this technique require transforming a dataset where items are of primitive types to a dataset of items of a user-defined type. The transformation logic is often defined in the problem or is part of the solution. We also create a comparator to order data items in the heap that is used to find the top k items.

If the problem statement or its solution follows the generic template below, it can be solved by applying the k-largest-items finding technique.

Example

Let's consider the following problem as an example to better understand how to identify and solve a problem using the custom compare technique.

Problem statement: Given an array of integers and an integer `k`, find the `k` most frequent integers.