Decision Tree Practice Lab Module 4

Course Q&A Machine Learning Specialization Advanced Learning Algorithms
1

Hello Mentors,

I’ve been working hard to complete this Decision Tree lab and wanted to share the steps I took in case you can help me figure out why Cell #7 keeps failing after submission.

  1. I followed the Coursera notebook step-by-step and used the provided hints to write code for each of the 4 graded exercises (compute_entropy, split_dataset, compute_information_gain, and get_best_split).
  2. I ran the test cells after each exercise individually, and all the tests passed (including the expected output from Cell #8 for entropy: 1.0).
  3. After completing all exercises, I clicked Restart and Run All to cleanly re-run the entire notebook in top-to-bottom order.
  4. I then submitted the notebook, but I’m still getting an error in Cell #7:
    “AssertionError: Entropy must be 1 with same amount of ones and zeros”
  5. I also saved the notebook to a folder and can upload it here if needed.

Would it help if I added print() statements inside the graded cells? I ran each one independently before the final restart and all the results looked correct.

Thank you for taking a look — I’m just trying to understand what’s causing this error even though everything works when I run it manually. I also used an updated notebook in the lab.

Best,
Mark

Screenshot Practice Lab Cell 7 Autograder error
Screenshot Practice Lab Cell 7 Autograder error841×910 104 KB

2

Passing the tests in the notebook does not prove your code is perfect. The notebook’s tests do not cover ever possible error.

The grader uses entirely different tests.

Your compute_entropy() function has a flaw.

3

Hi TMosh,

Thanks again for your help earlier. I’m still stuck on Cell 7 in the Decision Tree lab. When I run this manually:

print(compute_entropy(np.array([0, 1]))) # returns 1.0

…it prints 1.0 as expected.

But the grader still fails with:

AssertionError: Entropy must be 1 with same amount of ones and zeros

Based on your earlier suggestion and what I’ve seen in the forum I tried:

  • Flattening y at the start,
  • Returning 0.0 for empty input,
  • Returning 1.0 if y contains [0, 1] or [1, 0],
  • Making sure only one version of compute_entropy() exists,
  • Following the indentation tip in the hint section.

Really appreciate your time and guidance!

Best,
Mark

4

Please check your personal messages for instructions.