Exam DP-100 topic 3 question 28 discussion

Solution is wrong. Add dataset, Add split, Add Two Class, Add Permutation, Set Accuracy

According to the link https://gallery.azure.ai/Experiment/e2ccb5a5d9dc480489ba8ff0b7eb98ac correct answer - add data, split data, add two class, set accuracy, add permutation. You calculate the feature importance post you train your model and get the metrics.a

1) Add a dataset to the experiment. 2) Add a Split Data module to create training and test datasets. 3) Add a Two-Class Support Vector Machine module to initialize the SVM classifier. 4) Set the Metric for measuring performance property to Classification - Accuracy and then run the experiment. 5) Add a Permutation Feature Importance module and connect to the trained model and test dataset.

given answer is not correct: Add a dataset to the experiment Add a Split Data module to create training and test datasets Add a Two-Class Support Vector Machine module to initialize the SVM classifier Set the Metric for measuring performance property to Classification - Accuracy and then run the experiment Add a Permutation Feature Importance module and connect the trained model and test dataset

The provided answer is correct. Have a closer look at the doc provided: 1. Add the Permutation Feature Importance module to your experiment. .... 5. For metric for measuring performance, select ...

Add a dataset to the experiment ➡️ Add a split data module to create training and test datasets ➡️ Add a two-class support vector machine module to initialize the SVM classifier ➡️ Set the metric for measuring performance property classification - accuracy and then run the experiment ➡️ Add a permutation feature importance module and connect the trained model and test dataset.

4-5-1-2-3

Incorrect solution. Correct from ChatGPT is: Add a dataset to the experiment. Add a Split Data module to create training and test datasets. Add a Two-Class Support Vector Machine module to initialize the SVM classifier. Add a Permutation Feature Importance module and connect the trained model and test dataset. Set the Metric for measuring performance property to "Classification - Accuracy" and then run the experiment.

4-5-1-3-2

https://learn.microsoft.com/en-us/azure/machine-learning/component-reference/permutation-feature-importance#how-to-use-permutation-feature-importance

on exam 18-06-22

1. add dataset 2. split into train vs test 3. add model 4. connect everything 5. run experiment

There's a similar example on this page that explains the order properly: https://gallery.azure.ai/Experiment/e2ccb5a5d9dc480489ba8ff0b7eb98ac 1- Add the Adult Census Income Binary Classification dataset to your experiment. 2- Add a Split module to create training and test datasets. 3- Add a Two-Class Support Vector Machine module to initialize the SVM classifier. 4- Add a Train Model module to train the classifier, and connect the SVM module to the left input port and the training dataset to the right input port. Using the column selector set the Label column to income. 5- Add a Permutation Feature Importance module and connect the trained model and the test dataset to the left and right input ports respectively. Set the Metric for measuring 6- performance property to Classification - Accuracy. So the order seems to be, read data, split data, add Two-Class, add Permutation, set Accuracy

Solution appears to be correct as I tried it out in old ml studio, except that steps one and two can be reversed or be kept the as mentioned in the solution.

Answer is correct, Add Two Class, Add dataset, Add split, Add Permutation, Set Accuracy

on 19Oct2021

But according to this link https://docs.microsoft.com/en-us/azure/machine-learning/studio-module-reference/two-class-support-vector-machine, given answer is correct. Still confused because both the methods works fine, either you drag & drop Add two class or add dataset first.