Q11easyWhen is MAE (Mean Absolute Error) preferred over MSE (Mean Squared Error)?AWhen the dataset is empty.BWhen you want to punish large errors more.CWhen you want errors to be expressed in the same units as the target variable and reduce sensitivity to outliers.DWhen you are doing classification.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q12easyWhat is the Hold-out method?ATraining on a small dataset indefinitely.BUsing the same data to train and test.CSplitting a dataset into two distinct sets: one for training and one for evaluating.DTesting the model only on training data.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q13mediumAn e-mail spam filter must ensure that no legitimate mail is ever sent to spam. Which metric must be prioritized?APrecisionBAUCCRecallDAccuracyCheck AnswerReveal Answer & ExplanationDiscuss on Board
Q14easyIn regression, what does R-squared ($R^2$) indicate?AThe exact error value in currency.BThe number of training iterations.CThe number of features in the model.DThe proportion of variance in the target variable that is predictable from the features (Goodness-of-fit).Check AnswerReveal Answer & ExplanationDiscuss on Board
Q15mediumWhy is Accuracy a misleading metric for a dataset where 95% of samples belong to the "Negative" class?AA model can achieve 95% accuracy by simply predicting "Negative" for every single input.BAccuracy is only for regression models.CAccuracy does not work with numerical data.DAccuracy is too complex to calculate.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q16easyWhat is a Classification Report in Scikit-Learn?AA report about hardware health.BA summary of Precision, Recall, F1-Score, and Support for each class in the model.CA list of all rows in the test set.DA graph showing time usage.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q17easyYou perform 5-fold cross-validation. What do you do with the 5 different accuracy scores?APick the lowest one as the model's accuracy.BTake the mean (average) and standard deviation of all 5 scores.CPick the highest one as the model's accuracy.DDiscard the 4 worst ones.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q18easyWhat does the "Area Under the Curve" (AUC) represent in a binary classifier?AThe probability that the model will rank a randomly chosen positive instance higher than a randomly chosen negative one.BThe accuracy of the model.CThe area of the confusion matrix.DThe training time.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q19easyWhat is an ROC Curve plotting?APrecision vs. Training Data.BAccuracy vs. Time.CTrue Positive Rate vs. False Positive Rate.DBias vs. Variance.Check AnswerReveal Answer & ExplanationDiscuss on Board