You are asked to train a model for detecting fraudulent bank transactions. The bank studies show that the bank sustains more damage when a fraudulent transaction is approved compared to when a normal transaction is blocked. For the bank to sustain less damage, which cost matrix should you use to tune your model? (F stands for Fraudulent, and N stands for Normal.)

You have a dataset of 100,000 dog images comprised of 100 different breeds. If you want to use a Support Vector Machine (SVM) for a classification task, which approach should you choose and why?

Figure 2: Training Dataset for Pet Type Employed Age Debt College Degree Pet Yes 26 Yes No Cat Yes 45 No Yes Dog No 22 Yes Yes Dog No 50 Yes No Cat Yes 24 Yes No Cat No 33 Yes Yes Dog Yes 62 No No Cat Yes 39 Yes No Dog No 25 Yes No Dog   Review the table labeled Figure 2: Training Dataset for Pet Type. Assume we want to use a decision tree to predict a person’s pet type. Using misclassification error as the measure of node impurity, which attribute provides the best split?

You are asked to train a model for detecting cancer tumors in medical images. The hospital studies show that patients suffer more harm when a cancer tumor goes undiagnosed compared to when a benign tumor is treated for cancer. For patients to suffer less harm, which cost matrix should you use to tune your model? (F stands for Found Tumor, and N stands for No Tumor.)

Three classification models were trained on a dataset of previous medical images with the goal of predicting cancer cells. The first model (M1), which is complex, has a training error of 0.0012. The second model (M2), which is simpler, has a training error of 0.0021. The third model (M3), which is the simplest, has a training error of 0.1108. Which model would be the best one to use for classifying the medical images of future patients?

You have a dataset that contains information about different airplanes. For each airplane, more than 17 attributes, such as speed, capacity, price, brand, and so on, are available. You are asked to create a visual representation of this dataset for college students. What would NOT be a possible issue for learning models on this dataset?

Figure 7: Artificial Neural Network Dataset Instance 1 2 3 4 5 6 7 8 9 Actual Class 1 1 -1 1 1 -1 1 1 -1 Predicted Class 1 -1 -1 -1 -1 -1 -1 1 -1   Review the table labeled Figure 7: Artificial Neural Network Dataset. A forward pass of an Artificial Neural Network is finished and the data are classified in the table. What is the total sum of the squared errors?

Compute the cosine similarity and correlation coefficient between feature vectors X and Y: X = (1, 5, 2, 4) Y = (4, 4, 2, 6)

Figure 8: Dataset for Animal Type Color Continent Food Type Black Asia Meat Mammal Brown Europe Meat Mammal Black America Vegetables Mammal Green Africa Meat Mammal Black Asia Vegetables Mammal Black Africa Vegetables Bird Brown Europe Vegetables Bird Grey Asia Meat Mammal Brown America Meat Bird Green Asia Meat Mammal   Review the table labeled Figure 8: Dataset for Animal Type. You trained a Naive Bayes Classifier on the dataset. If a record contains the values (Black, Africa, Vegetables), what Type would the model predict?

Select all possible columns that can be set as class label for a data mining task.   Tid Refund marital Status Taxable Income Cheat 1 Yes Single 125k No 2 No Married 100k No 3 No Single 70k No 4 Yes Married 120k No 5 No Divorced 95k Yes 6 No Married 60k No 7 Yes Divorced 220k No 8 No Single 95k Yes 9 No Married 75k No 10 No Single 90k Yes