Answer all $!!$ A$.$ Ensemble methods help in $\_\_\_\_\_\_\_\_\_\_.$a$.$ Improving predictionsb. Increasing biasc. Decreasing varianced. Decreasing biasB. $\_\_\_\_\_\_\_\_\_\_\_\_\_$ is an ensemble method that is used for sequential learning as it exploits the dependency between models by giving higher weights to mislabelled examples.a$.$XGBoostb. Random forestc. Gradient boostd. AdaBoostC. $\_\_\_\_\_\_\_\_\_\_\_$ was introduced to deal with the limitations of decision trees.a$.$ Boostingb. Bootstrappingc. Baggingd. AveragingD. Boosting is disadvantageous because it is:a$.$Difficult to implement in real$-$timeb.A resilient methodc.Difficult to scale upd.Sensitive to outliersE. Base models in the stacking architecture can include:a$.$Linear modelsb.Support vector machinesc.Non$-$linear modelsd.Neural networksF. The default index that is used by random forest classifiers to decide which features are important is:a$.$HashMapsb. Ginic. BTreed. BitMapG.AdaBoost$$will focus on a weak learner such as$$a decision tree$$with only one split called$\_\_\_\_\_\_\_\_\_\_.$a$.$ Branchb. Decision stumpc. Noded. LeafH. A stump can use $\_\_\_\_\_\_\_\_\_\_$to make a decision.a$.$Two or more variablesb.Only one variablec.All the variablesd.One or more variablesI. Gradient boosting classifiers use the $\_\_\_\_\_\_\_\_\_\_$ algorithm at each stage to minimise errors.a$.$Gradient ascentb.Gradient descentc.Gradient optimisingd.Stochastic gradient ascentJ. At present,$\_\_\_\_\_\_\_\_\_\_$ is the de facto standard algorithm for getting accurate results from predictive$$modelling$$with$$machine learning.a$.$AdaBoostb.RFc$.$KNNd$.$XGBoost