1. Examples of Classification applications include A Profiling B Co-occurrence grouping. Clustering D None of the above. 2 Errors in classification models A Are measured in a similar way to regression, except not with squares B Can only be determined when the outcomes are binary C Are more complicated than in value estimation models D. All of the above. 3. Applying multiple regression to classification presents challenges because A The outcome is binary B Only gives estimates of the probability of being in a group (class) C Can give probability outcomes that are not between 0 and 1. D All of the above Logistic regression is preferred to linear regression because A It is logical rather than linear Bit gives probability estimates between 0 and 1 C Linear regression cannot accommodate binary predictor (independent variables D. All of the above 5. A discriminantfunction Als a straightline B Always divides two groups perfectly Cls a function that separates the majority of values into two groups Dis the same as cluster analysis 6. KNN stands for A Knowledge that's Non-Numeric B Knowledge using Naive Numbers CK Nearest Neighbours D None of the above 7. KNN is intuitively appealing but presents practical challenges, including A How do you measure "similarity"? B How do you measure similarity of categorical attributes? C Are all attributes of the same importance? D Allof the above 8 KNN differs from logistic regression and discriminant functions in that A There is no formula to determine which group a case belongs to B You must store all the past data C Neither a norb is true D Botha and b are true 9 Methods that estimate the probability or likelihood that a case belongs to a group are incomplete without A A causal analysis that explains the contribution of each variable BA decision rule to how to assign cases to groups based upon the probability CA measure of error of the probability Nothing else is needed since you assign the case to the group if the probability exceeds 50% 10. Among methods for binary dassification A Logistic regression is the most accurate, B Ducriminant functions work best CKNN is the method of last resort D They all work reasonably well