Course 3: Week 1 (Lab 2) problem

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

I was curious about last part of Lab that how can I create plot as It was created in upper part of code.

p1
p11366×407 24.8 KB

I want to create a plot as this:
p3
p31366×437 42.8 KB

but facing error like this:
Error: "ValueError Traceback (most recent call last)
in
3
4 # Visualize the fit
----> 5 visualize_fit(X_val_high, mu_high, var_high)
6
7 # Draw a red circle around those outliers

~/work/utils.py in visualize_fit(X, mu, var)
44
45 X1, X2 = np.meshgrid(np.arange(0, 35.5, 0.5), np.arange(0, 35.5, 0.5))
—> 46 Z = multivariate_gaussian(np.stack([X1.ravel(), X2.ravel()], axis=1), mu, var)
47 Z = Z.reshape(X1.shape)
48

~/work/utils.py in multivariate_gaussian(X, mu, var)
30 var = np.diag(var)
31
—> 32 X = X - mu
33 p = (2* np.pi)(-k/2) * np.linalg.det(var)(-0.5) *
34 np.exp(-0.5 * np.sum(np.matmul(X, np.linalg.pinv(var)) * X, axis=1))

ValueError: operands could not be broadcast together with shapes (5041,2) (11,) "

Can anyone help to solve this error. Thankyou

regards
Ahmad Fareed Khan

2

Since the larger dataset has 11 features, you can’t make a 2D plot of it.

3

so, how can we make the plot(or any other kind of view) for lots of features like in real world examples for analysis? Is there any other method for that?

4

Generally you can’t plot a multi-dimensional data set. There is a plot in this assignment only so you can see how the technique works.

5

Thanks! Tom for guidance.