Distance computation (distance.cdist) in SciPy - Cheat Sheet & Quick Revision

Recall & Review

beginner

What does the function scipy.spatial.distance.cdist compute?

cdist computes the distance between each pair of points from two collections of points. It returns a matrix where each element is the distance between a point in the first set and a point in the second set.

beginner

What are the inputs required by distance.cdist?

Two arrays of points (2D arrays), where each row is a point and each column is a dimension. Optionally, a metric string to specify the type of distance (like 'euclidean', 'cityblock').

beginner

Name two common distance metrics you can use with cdist.

Euclidean distance (straight line) and Manhattan distance (sum of absolute differences).

intermediate

How does cdist differ from computing distances in a loop?

cdist is optimized and vectorized, so it is much faster and simpler than manually looping over points to compute distances.

intermediate

What shape does the output of cdist have if the first input has shape (m, n) and the second input has shape (p, n)?

The output shape is (m, p), where each element [i, j] is the distance between the i-th point in the first set and the j-th point in the second set.

What does scipy.spatial.distance.cdist return?

A A single distance value between two points

B A matrix of distances between points in two sets

C A list of points sorted by distance

D A boolean matrix indicating if points are equal

cdist returns a matrix where each element is the distance between a point in the first set and a point in the second set.

Which of these is NOT a valid metric for cdist?

A euclidean

B cityblock

C random

D cosine

'random' is not a valid distance metric for cdist. Common metrics include 'euclidean', 'cityblock', and 'cosine'.

If you have 5 points in set A and 3 points in set B, what will be the shape of the output from cdist(A, B)?

A (1, 15)

B (3, 5)

C (8, 1)

D (5, 3)

The output shape is (number of points in A, number of points in B), so (5, 3).

Why is cdist preferred over manual loops for distance calculation?

A It is faster and simpler to use

B It only works for 1D points

C It returns distances as strings

D It uses less memory

cdist is optimized and vectorized, making it faster and easier than manual loops.

Which Python library provides the cdist function?

A scipy

B pandas

C numpy

D matplotlib

cdist is part of scipy.spatial.distance.

Explain how you would use scipy.spatial.distance.cdist to find distances between two sets of points.

Think about inputs, metric, and output shape.

Describe the difference between Euclidean and Manhattan distance metrics used in cdist.

Imagine walking in a city grid vs flying straight.