Page 318 - AI Computer 10

Step 10: Build the model by clicking the Test and Score option under the Evaluate widget. After that,
connect the Image Embedding icon with the Test and Score icon.
Step 11: Select different algorithms for classification such as Logistic Regression, SVM, and Random
Forest under the Model widget. Connect these algorithms icons with the Test and Score icon
to check the performance of each.

Step 12: Double-click on Test and Score icon to view the evaluation metric like Accuracy, F1 Score,
Precision, and Recall for all the three algorithms. Here, Logistic Regression gives the best
accuracy.
Step 13: Drag the Confusion Matrix icon onto the Canvas and connect it with the Test and Score icon.

Step 14: Double-click the Confusion Matrix icon to view the distribution of correct and incorrect
predictions.

Step 15: Observe correct and incorrect predictions for each algorithm. Now, upload the Testing dataset,
similar to the Training dataset, by dragging the Import Images icon onto the Canvas.

After that, add Image Viewer icon and Image Embedding icon on the Canvas and connect both
of them with the Test Data icon.
Step 16: Drag Predictions icon from the Evaluate widget on to the Canvas and connect it with the
Image Embedding icon.

Step 17: Click on Logistic Regression and connect it to Predictions. Also, connect Image Embedding (of
Training Data) to Logistic Regression as shown.

184
184