Q1easyIf your model has High Bias, what should you do?AUse fewer data samples.BRemove all features.CIncrease model complexity (add more features, use a more complex algorithm).DSimplify the model.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q2easyWhich algorithm is a "non-parametric" supervised learning method?ALogistic RegressionBK-Nearest Neighbors (KNN)CLinear RegressionDPolynomial RegressionCheck AnswerReveal Answer & ExplanationDiscuss on Board
Q3easyWhat is the main disadvantage of K-Nearest Neighbors (KNN) when used on large datasets?AIt requires a long training phase.BIt cannot handle numerical data.CIt is computationally expensive during the prediction phase, as it must calculate the distance to all training points.DIt only works on binary classification.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q4easyWhy is Hold-out Validation (train/test split) necessary?ATo save memory.BTo speed up training.CTo ensure the model can generalize to new data, rather than just memorizing the training data.DTo make the algorithm easier to debug.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q5easyIn Scikit-Learn, what is the role of train_test_split?AIt trains the model twice.BIt sorts the data by label.CIt duplicates the training data.DIt splits the dataset into training and testing portions (80% train, 20% test).Check AnswerReveal Answer & ExplanationDiscuss on Board
Q6easyWhat is the Kernel Trick in SVMs?AA way to hack the kernel of the OS.BA method for compressing data. data_split.pyCAn error handling mechanism.DA method to project non-linear data into a higher-dimensional space where it can be separated by a linear hyperplane.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q7hardWhich algorithm is most commonly associated with Recursive Partitioning?ALinear RegressionBPCACK-MeansDDecision TreeCheck AnswerReveal Answer & ExplanationDiscuss on Board
Q8easyWhat is Regularization in linear models?AAdding a penalty term to the loss function to prevent coefficients from becoming too large, thereby reducing overfitting.BMaking the model faster.CChanging the learning rate.DAutomatically cleaning the input data.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q9easyWhat is Support Vector Machine (SVM) trying to achieve?ATo find the median value.BTo generate new training data.CTo minimize the number of features.DTo find the hyperplane that maximizes the margin (distance) between different classes.Check AnswerReveal Answer & ExplanationDiscuss on Board
Q10easyWhat is the difference between Regression and Classification metrics?AClassification always uses RMSE.BThere is no difference.CRegression uses accuracy; Classification uses MSE.DRegression uses metrics like MSE/MAE (measuring error magnitude); Classification uses Accuracy/F1 (measuring correct/incorrect category labels).Check AnswerReveal Answer & ExplanationDiscuss on Board