ISYE 6501 - Final Prep – Midterm Questions With Correct Answers

ISYE 6501 - Final Prep – Midterm Questions With Correct Answers Types of classification models KNN SVM Types of Clustering Models K-Means Clustering Types of Response Prediction Models ARIMA CART Exponential Smoothing Linear Regression Logistic Regression Random Forest Types of Validation Method Cross Validation Types of Variation Estimate Models GARCH Models that use Time Series Data ARIMA CUSUM GARCH Exponential Smoothing Do models tend to perform better or worse on test sets? Worse, because models are fit to the training data so they'll likely perform worse on the test set. How are margins defined in SVMs? The length of the distance to the closest point (e.g., if points lie closer to the separation plane then the margin is narrower) What will happen if you need to move the margin of an SVM? The model will perform worse on training and test data. This is done when the cost of misclassifying something may be sever (e.g., misclassifying poisonous mushroom as safe) What happens if multiple models have similar R^2? They'll all perform similarly on test data (I think?) When is a model autoregressive? When it uses information from the previous reading (t-1). No