C3_w1_anomaly detection

Course Q&A Machine Learning Specialization Unsupervised Learning, Recommenders, Reinforcement
1

Tried my version of the code to solve Exercise 2 (setting the threshold epsilon to get max F1 score)…after multiple attempts i ended up using the code given in the hints section to no avail!!

please help!

UNQ_C2

GRADED FUNCTION: select_threshold

def select_threshold(y_val, p_val): 
    """
    Finds the best threshold to use for selecting outliers 
    based on the results from a validation set (p_val) 
    and the ground truth (y_val)
    
    Args:
        y_val (ndarray): Ground truth on validation set
        p_val (ndarray): Results on validation set
        
    Returns:
        epsilon (float): Threshold chosen 
        F1 (float):      F1 score by choosing epsilon as threshold
    """ 

    best_epsilon = 0
    best_F1 = 0
    F1 = 0
    
    step_size = (max(p_val) - min(p_val)) / 1000
    
    for epsilon in np.arange(min(p_val),max(p_val),step_size):
        
   # moderator edit: code removed

        
    return best_epsilon, best_F1
2

Hi @Raghavendran_n ,

The ‘if’ statement is outside of the ‘for’ loop, so ‘if’ block of code only get executed once.

3

Thank you, KIc !! spent hours figuring this out and it turns out to be bad indentation!

thanks, again!

4

Please do not post your code on the forum. That’s not allowed by the community standard.

The mentors are not here to fix your code - we’re here to give you hints about how you can fix your code yourself.

5

Can some one help me on this, I’m getting below error but all the testcases passed getting compile error

# UNQ_C1
# GRADED FUNCTION: estimate_gaussian
# mentor edit: code removed

NameError Traceback (most recent call last)
in
1 # Find the outliers in the training set
----> 2 outliers = p < epsilon
3
4 # Visualize the fit
5 visualize_fit(X_train, mu, var)

NameError: name 'p' is not defined

{mentor edit: image and code removed}

NameError Traceback (most recent call last)
in
1 p_val = multivariate_gaussian(X_val, mu, var)
----> 2 epsilon, F1 = select_threshold(y_val, p_val)
3
4 print('Best epsilon found using cross-validation: e' epsilon)
5 print('Best F1 on Cross Validation Set: f' F1)

NameError: name ‘select_threshold’ is not defined

6

Please do not share your code on the forum. That’s not allowed by the Code of Conduct.

I have edited your post to remove the code.

Posting the error messages is fine.

1 Like
7

Tip: Be sure that every time you open a notebook, you run all of the cells starting from the top of the page.

8

yes i have run all the cells, only compile error, not sure what was the issue

9

The variable ‘p’ is defined when you run this cell:

image
image699×213 7.88 KB

It’s in the Exercise 1 part of the notebook. The error you show is way down at the bottom of Exercise 2.

That;s why I mentioned that you have to run all of the cells every time you open a notebook, starting right at the top.

10

Hello

Find the outliers in the training set

outliers = p < epsilon
my error is displaying first block then the expected outcome cell

later the part which you mentioning

Returns the density of the multivariate normal

at each data point (row) of X_train

p = multivariate_gaussian(X_train, mu, var) is coming

11

I am struggling to understand your question.

Please post a screen capture image that shows whatever error messages you are seeing.

Avoid using text copy-and-paste, because it’s being rendered as Markdown, which is very confusing.

12

image
image706×628 107 KB

I have run all the cells, code has been written in Exercise 1 , after the code block i can see another which code has already been defined, post executing that block i can see the error, same for the Exercise 2 also.
and P defining this all in the next cell blocks.

image
image736×625 101 KB

13

Sorry, I don’t understand your description of the situation for “name p is not defined”.

For the 2nd image you posted, it appears that function has an indentation error.