What modification can be made to bubble sort to make it perform better on partially sorted datasets?

The modification that can be made to the Bubble Sort algorithm to make it perform better on partially sorted datasets is the introduction of a flag to check if a swap has occurred in a particular pass. This modification is often referred to as the "Optimized Bubble Sort" or the "Adaptive Bubble Sort".

Here are the steps:

1. Start the outer loop, which iterates over each element in the array.

2. Before the inner loop begins, set a boolean flag to false. This flag is used to check if any swapping occurred in the inner loop.

3. The inner loop starts from the first element and compares each pair of elements, swapping them if they are in the wrong order.

4. If a swap has occurred in the inner loop, set the flag to true.

5. After each pass of the inner loop, check the flag. If it remains false, this means no swap has occurred in the last pass, indicating that the array is already sorted. In this case, break the outer loop.

6. If the flag is true, reset it to false and continue with the next pass.

This modification improves the Bubble Sort's performance on partially sorted datasets because it avoids unnecessary passes once the array is sorted.