{
  "title": "Bagging Mastery: 100 MCQs",
  "description": "A comprehensive set of multiple-choice questions designed to test and deepen your understanding of Bagging (Bootstrap Aggregating), starting with easy-level concepts (1–30).",
  "questions": [
    {
      "id": 1,
      "questionText": "What does Bagging stand for in ensemble learning?",
      "options": [
        "Bootstrap Aggregating",
        "Bayesian Aggregation",
        "Binary Aggregation",
        "Batch Averaging"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Bagging stands for Bootstrap Aggregating. It improves model stability and accuracy by training multiple models on random subsets of the dataset and aggregating their predictions."
    },
    {
      "id": 2,
      "questionText": "What is the main purpose of Bagging?",
      "options": [
        "Increase complexity",
        "Normalize data",
        "Reduce variance",
        "Reduce bias"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging reduces variance by averaging predictions from multiple models trained on different bootstrap samples, helping improve model stability."
    },
    {
      "id": 3,
      "questionText": "Which type of models is Bagging most effective with?",
      "options": [
        "Clustering models",
        "Linear models only",
        "High variance models",
        "High bias models"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging is especially effective with high variance models (like decision trees) because averaging multiple models reduces variance and prevents overfitting."
    },
    {
      "id": 4,
      "questionText": "How are the datasets generated in Bagging?",
      "options": [
        "By splitting features into groups",
        "By normalizing the original dataset",
        "By removing outliers",
        "By randomly sampling with replacement"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Bagging uses bootstrap sampling, which randomly selects samples with replacement to create multiple training datasets for each model in the ensemble."
    },
    {
      "id": 5,
      "questionText": "In Bagging, how is the final prediction typically made?",
      "options": [
        "By using the last trained model only",
        "By averaging or majority voting",
        "By multiplying predictions",
        "By choosing the first model’s output"
      ],
      "correctAnswerIndex": 1,
      "explanation": "The final prediction in Bagging is usually made by averaging the outputs for regression tasks or majority voting for classification tasks."
    },
    {
      "id": 6,
      "questionText": "Which of the following is NOT a benefit of Bagging?",
      "options": [
        "Reduces overfitting",
        "Improves prediction stability",
        "Reduces bias significantly",
        "Reduces variance"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging primarily reduces variance. It may slightly reduce bias, but it does not significantly reduce bias. Other ensemble methods like boosting are better for bias reduction."
    },
    {
      "id": 7,
      "questionText": "Which algorithm is commonly used with Bagging?",
      "options": [
        "Naive Bayes",
        "Linear Regression",
        "Decision Trees",
        "K-Means"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Decision Trees are commonly used with Bagging because they have high variance, and Bagging reduces this variance effectively."
    },
    {
      "id": 8,
      "questionText": "What is the main difference between Bagging and a single model?",
      "options": [
        "Bagging uses multiple models to reduce variance",
        "Bagging uses only one model",
        "Bagging removes all data randomness",
        "Bagging increases overfitting intentionally"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Bagging trains multiple models on different random subsets and aggregates their predictions, which reduces variance compared to a single model."
    },
    {
      "id": 9,
      "questionText": "Bootstrap samples in Bagging are:",
      "options": [
        "Selected based on feature importance",
        "Always smaller than 10% of dataset",
        "Randomly drawn with replacement",
        "Drawn without replacement"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bootstrap sampling involves selecting data points randomly with replacement, allowing some points to appear multiple times in a sample."
    },
    {
      "id": 10,
      "questionText": "Bagging is mainly used for which type of problem?",
      "options": [
        "Only clustering",
        "Only anomaly detection",
        "Only dimensionality reduction",
        "Classification and regression"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Bagging is an ensemble method applicable to both classification and regression tasks."
    },
    {
      "id": 11,
      "questionText": "In Bagging, increasing the number of models generally:",
      "options": [
        "Increases bias",
        "Makes individual models more complex",
        "Reduces variance and improves stability",
        "Reduces dataset size"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Adding more models to Bagging averages predictions over more models, reducing variance and improving prediction stability."
    },
    {
      "id": 12,
      "questionText": "Which ensemble method uses boosting instead of averaging?",
      "options": [
        "Random Forest",
        "Bagging",
        "Boosting",
        "K-Means"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Boosting is an ensemble method that sequentially trains models, focusing on errors of previous models, rather than averaging independent models like Bagging."
    },
    {
      "id": 13,
      "questionText": "Why does Bagging reduce overfitting in high variance models?",
      "options": [
        "Because it removes data noise",
        "Because it increases bias",
        "Because it averages multiple models’ predictions",
        "Because it uses fewer features"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging reduces overfitting by training multiple models on different samples and averaging their predictions, which stabilizes the output."
    },
    {
      "id": 14,
      "questionText": "Random Forest is a type of:",
      "options": [
        "Bagging with feature randomness",
        "Boosting with weighting",
        "Single decision tree",
        "Dimensionality reduction method"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Random Forest is an extension of Bagging where trees are trained on bootstrap samples and each split considers a random subset of features to reduce correlation among trees."
    },
    {
      "id": 15,
      "questionText": "Which of these is a key requirement for Bagging to be effective?",
      "options": [
        "High variance of base models",
        "High bias of base models",
        "Small dataset size",
        "Single feature only"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Bagging is most effective when base models have high variance; averaging their outputs reduces variance and stabilizes predictions."
    },
    {
      "id": 16,
      "questionText": "Bagging works best with:",
      "options": [
        "Stable learners like linear regression",
        "Clustering models",
        "Unstable learners like decision trees",
        "Dimensionality reduction models"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging reduces variance. Unstable learners with high variance benefit the most, while stable learners like linear regression do not gain much."
    },
    {
      "id": 17,
      "questionText": "How is the randomness introduced in Bagging?",
      "options": [
        "Through normalization",
        "Through adding noise to labels",
        "Through bootstrap sampling of data",
        "Through reducing feature space"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Randomness in Bagging comes from creating multiple bootstrap samples from the original dataset."
    },
    {
      "id": 18,
      "questionText": "In classification with Bagging, the final class is decided by:",
      "options": [
        "Multiplying probabilities",
        "Weighted averaging",
        "Majority voting",
        "Selecting first model output"
      ],
      "correctAnswerIndex": 2,
      "explanation": "For classification, Bagging predicts the class that receives the most votes among all models."
    },
    {
      "id": 19,
      "questionText": "Which of the following statements is TRUE about Bagging?",
      "options": [
        "It decreases dataset size",
        "It reduces variance without greatly affecting bias",
        "It increases variance",
        "It is only used for regression"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Bagging reduces variance by averaging predictions, while bias remains mostly unchanged."
    },
    {
      "id": 20,
      "questionText": "Bagging can be used with which base learners?",
      "options": [
        "Only decision trees",
        "Only clustering models",
        "Any model that benefits from variance reduction",
        "Only linear models"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Any high-variance model can benefit from Bagging, not just decision trees."
    },
    {
      "id": 21,
      "questionText": "Bootstrap samples are the same size as the original dataset. True or False?",
      "options": [
        "False",
        "Depends on the algorithm",
        "True",
        "Depends on features"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Typically, each bootstrap sample has the same number of instances as the original dataset but is sampled with replacement."
    },
    {
      "id": 22,
      "questionText": "Which scenario is ideal for using Bagging?",
      "options": [
        "Small datasets with no noise",
        "Low variance models",
        "High variance models prone to overfitting",
        "Single-feature linear regression"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging helps reduce overfitting in high variance models by averaging predictions from multiple models."
    },
    {
      "id": 23,
      "questionText": "Bagging helps in prediction stability by:",
      "options": [
        "Reducing dataset size",
        "Changing the loss function",
        "Increasing model depth",
        "Reducing fluctuations due to individual models"
      ],
      "correctAnswerIndex": 3,
      "explanation": "By averaging multiple models, Bagging reduces the impact of fluctuations from any single model, improving stability."
    },
    {
      "id": 24,
      "questionText": "Which of these is an ensemble learning technique like Bagging?",
      "options": [
        "Boosting",
        "PCA",
        "Feature Scaling",
        "K-Means"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Boosting is another ensemble learning technique that differs from Bagging by sequentially training models."
    },
    {
      "id": 25,
      "questionText": "Does Bagging always improve model performance?",
      "options": [
        "It only works with linear models",
        "It decreases performance for high variance models",
        "It improves performance if the base model is high variance",
        "It always improves performance"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging improves performance primarily for models with high variance; stable models may not gain significant improvement."
    },
    {
      "id": 26,
      "questionText": "In Bagging, can the same instance appear multiple times in a bootstrap sample?",
      "options": [
        "Yes, due to sampling with replacement",
        "No, each instance appears only once",
        "Only if dataset is small",
        "Depends on features"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Bootstrap sampling is done with replacement, so some instances may appear multiple times in the same sample."
    },
    {
      "id": 27,
      "questionText": "Bagging reduces overfitting by:",
      "options": [
        "Adding noise to data",
        "Increasing learning rate",
        "Reducing feature dimensionality",
        "Averaging multiple models trained on different data"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Averaging predictions from multiple models trained on bootstrap samples reduces overfitting and variance."
    },
    {
      "id": 28,
      "questionText": "Which statement is TRUE about Random Forest compared to Bagging?",
      "options": [
        "Random Forest uses only one tree",
        "Random Forest adds feature randomness to Bagging",
        "Random Forest does not use bootstrap sampling",
        "Random Forest uses boosting"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Random Forest is Bagging with additional feature randomness at each split to decorrelate trees."
    },
    {
      "id": 29,
      "questionText": "Which error does Bagging aim to reduce the most?",
      "options": [
        "Feature selection error",
        "Variance",
        "Irreducible error",
        "Bias"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Bagging primarily reduces variance in high-variance models, leading to more stable predictions."
    },
    {
      "id": 30,
      "questionText": "Which type of datasets benefit most from Bagging?",
      "options": [
        "Datasets meant for clustering",
        "Small, perfectly clean datasets",
        "Large datasets with noisy labels",
        "Datasets with single features only"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging is especially useful for large datasets with noisy labels or high variance models to stabilize predictions."
    },
    {
      "id": 31,
      "questionText": "What is the role of the number of estimators (trees) in Bagging?",
      "options": [
        "Increasing it increases bias",
        "It controls feature selection",
        "Increasing it generally reduces variance",
        "It reduces dataset size"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Increasing the number of base models (trees) in Bagging helps in averaging more predictions, which reduces variance and stabilizes the model."
    },
    {
      "id": 32,
      "questionText": "When performing regression with Bagging, which aggregation method is used?",
      "options": [
        "Majority voting",
        "Averaging predictions",
        "Multiplying predictions",
        "Weighted voting"
      ],
      "correctAnswerIndex": 1,
      "explanation": "For regression, Bagging combines predictions by averaging outputs from all models."
    },
    {
      "id": 33,
      "questionText": "Which hyperparameter of base models impacts Bagging performance the most?",
      "options": [
        "Learning rate",
        "Model depth (for decision trees)",
        "Kernel type",
        "Number of classes"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Decision tree depth influences individual model variance. Deep trees are high variance and benefit most from Bagging."
    },
    {
      "id": 34,
      "questionText": "If a Bagging ensemble is underfitting, which approach can help?",
      "options": [
        "Decrease features",
        "Reduce sample size",
        "Reduce number of trees",
        "Increase base model complexity"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Underfitting occurs when models are too simple. Increasing the complexity of base models allows each to capture more patterns, improving ensemble performance."
    },
    {
      "id": 35,
      "questionText": "Bagging can help reduce overfitting caused by:",
      "options": [
        "Irreducible error",
        "Small dataset size",
        "High bias in base learners",
        "High variance in base learners"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Bagging reduces overfitting that arises from high variance models by averaging multiple models trained on bootstrap samples."
    },
    {
      "id": 36,
      "questionText": "How does Bagging impact the training time?",
      "options": [
        "Has no effect",
        "Increases training time linearly with number of models",
        "Decreases training time",
        "Reduces only for regression"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Training multiple models increases computational cost, as each model is trained separately on a bootstrap sample."
    },
    {
      "id": 37,
      "questionText": "Which metric would best evaluate Bagging for classification?",
      "options": [
        "Silhouette Score",
        "Mean Squared Error",
        "Accuracy, F1-score, or AUC",
        "R-squared"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Accuracy, F1-score, and AUC are standard metrics for classification, suitable for evaluating Bagging ensembles."
    },
    {
      "id": 38,
      "questionText": "Bagging helps in scenarios where the model is:",
      "options": [
        "High variance but low bias",
        "Low bias and low variance",
        "Unsupervised",
        "High bias but low variance"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Bagging is most beneficial for high variance models; it averages predictions to reduce variance while bias remains low."
    },
    {
      "id": 39,
      "questionText": "If bootstrap samples are too small, what is likely to happen?",
      "options": [
        "Bias decreases",
        "Variance reduction decreases",
        "Model becomes unsupervised",
        "Training time increases"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Smaller bootstrap samples provide less diversity and reduce the effectiveness of variance reduction in Bagging."
    },
    {
      "id": 40,
      "questionText": "Bagging can be combined with which technique for better performance?",
      "options": [
        "Normalization only",
        "PCA without ensemble",
        "Single linear regression",
        "Random feature selection (Random Forest)"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Combining Bagging with random feature selection, as in Random Forests, further decorrelates trees and improves performance."
    },
    {
      "id": 41,
      "questionText": "Which of the following is true about Bagging and bias?",
      "options": [
        "Bias is irrelevant",
        "Bias may remain mostly unchanged",
        "Bias increases significantly",
        "Bias is always reduced"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Bagging primarily reduces variance. Bias generally remains the same because base learners are not modified."
    },
    {
      "id": 42,
      "questionText": "In Bagging, how are outliers in the training data handled?",
      "options": [
        "They are removed automatically",
        "They cause model to ignore majority classes",
        "They have no effect",
        "They are partially mitigated by averaging predictions"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Outliers may affect individual models, but averaging predictions reduces their impact on final output."
    },
    {
      "id": 43,
      "questionText": "Bagging with deep trees is preferred over shallow trees because:",
      "options": [
        "Shallow trees overfit more",
        "Shallow trees have high variance",
        "Deep trees reduce bias automatically",
        "Deep trees have higher variance which Bagging reduces"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Bagging reduces variance. Deep trees tend to overfit (high variance), so Bagging stabilizes them."
    },
    {
      "id": 44,
      "questionText": "Which is an advantage of Bagging over a single model?",
      "options": [
        "Faster training",
        "Improved prediction stability",
        "Automatic feature selection",
        "Reduced number of features"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Bagging improves stability and reduces variance by averaging predictions from multiple models."
    },
    {
      "id": 45,
      "questionText": "Bagging can help in which real-world scenario?",
      "options": [
        "Single linear regression on clean data",
        "Unsupervised clustering",
        "Classifying perfectly separable data",
        "Predicting stock prices with high-variance trees"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Bagging is useful in high-variance prediction problems, such as stock price prediction with complex decision trees."
    },
    {
      "id": 46,
      "questionText": "Why might Bagging not improve a linear regression model?",
      "options": [
        "Linear regression is unstable",
        "Bagging cannot be used for regression",
        "It always decreases performance",
        "Linear regression is a low variance model"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Linear regression is a stable, low-variance model. Bagging does not significantly improve performance in such cases."
    },
    {
      "id": 47,
      "questionText": "In Bagging, increasing correlation among base models:",
      "options": [
        "Improves variance reduction",
        "Decreases bias automatically",
        "Does not matter",
        "Reduces ensemble effectiveness"
      ],
      "correctAnswerIndex": 3,
      "explanation": "High correlation among base models reduces the benefit of averaging, making Bagging less effective."
    },
    {
      "id": 48,
      "questionText": "When using Bagging, what should you do to reduce correlation among trees?",
      "options": [
        "Use random subsets of features (Random Forest approach)",
        "Use fewer trees",
        "Increase bootstrap sample size",
        "Use shallow trees only"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Randomly selecting features at each split reduces correlation among trees, enhancing Bagging effectiveness."
    },
    {
      "id": 49,
      "questionText": "Which is true about Bagging in small datasets?",
      "options": [
        "It always works perfectly",
        "It may not improve performance much",
        "It increases model complexity",
        "It reduces bias significantly"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Bagging relies on diverse bootstrap samples. In small datasets, diversity is limited, reducing its effectiveness."
    },
    {
      "id": 50,
      "questionText": "What is the effect of Bagging on model interpretability?",
      "options": [
        "Interpretability increases",
        "Interpretability decreases compared to single model",
        "It simplifies decision trees",
        "No effect"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Ensembling multiple models makes it harder to interpret predictions compared to a single model."
    },
    {
      "id": 51,
      "questionText": "Which combination is commonly used in practice?",
      "options": [
        "Bagging with linear regression on clean data",
        "Bagging with decision trees (Random Forest)",
        "Bagging with K-Means",
        "Bagging with PCA"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Bagging with decision trees, as in Random Forests, is the most common and effective practical implementation."
    },
    {
      "id": 52,
      "questionText": "What is the effect of increasing the number of trees beyond a certain point?",
      "options": [
        "Training time decreases",
        "Overfitting increases",
        "Bias increases",
        "Variance reduction saturates"
      ],
      "correctAnswerIndex": 3,
      "explanation": "After a certain number of trees, adding more provides little additional variance reduction, but training cost increases."
    },
    {
      "id": 53,
      "questionText": "Bagging is more suitable than boosting when:",
      "options": [
        "High variance base learners need stabilization",
        "High bias learners need improvement",
        "The dataset is very small",
        "Features are unimportant"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Bagging reduces variance, while boosting is more focused on reducing bias and sequential learning."
    },
    {
      "id": 54,
      "questionText": "What type of error does Bagging primarily address?",
      "options": [
        "Feature error",
        "Bias",
        "Irreducible error",
        "Variance"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Bagging reduces variance errors by averaging predictions from multiple models."
    },
    {
      "id": 55,
      "questionText": "How can Bagging handle noisy labels?",
      "options": [
        "It removes noisy labels automatically",
        "Noise increases ensemble variance",
        "Averaging reduces the effect of noisy instances",
        "Noise has no effect"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Averaging predictions from multiple models reduces the influence of noise in individual training samples."
    },
    {
      "id": 56,
      "questionText": "In Random Forest, what differentiates it from plain Bagging?",
      "options": [
        "Sequential learning",
        "Random feature selection at each split",
        "Boosting weights",
        "No bootstrap sampling"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Random Forest introduces feature randomness at each split in addition to Bagging to decorrelate trees."
    },
    {
      "id": 57,
      "questionText": "Bagging ensemble predictions are robust because:",
      "options": [
        "Only the first model matters",
        "It reduces bias completely",
        "All models use the same data",
        "Individual model errors cancel out"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Averaging predictions from diverse models helps cancel out individual errors, leading to more robust outputs."
    },
    {
      "id": 58,
      "questionText": "Which is NOT a hyperparameter of Bagging?",
      "options": [
        "Base model type",
        "Learning rate",
        "Bootstrap sample size",
        "Number of estimators"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Learning rate is not a hyperparameter for Bagging; it is used in boosting algorithms."
    },
    {
      "id": 59,
      "questionText": "How does Bagging affect overfitting on noisy datasets?",
      "options": [
        "Does not affect overfitting",
        "Increases overfitting",
        "Reduces overfitting",
        "Eliminates bias completely"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Averaging predictions reduces variance, which helps in reducing overfitting on noisy datasets."
    },
    {
      "id": 60,
      "questionText": "Bagging works best when base models are:",
      "options": [
        "Stable and low variance",
        "Unstable and high variance",
        "Linear regression only",
        "Perfectly accurate"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Bagging reduces variance, so it works best with unstable, high-variance models like decision trees."
    },
    {
      "id": 61,
      "questionText": "Increasing diversity among base learners in Bagging:",
      "options": [
        "Reduces stability",
        "Increases bias",
        "Improves ensemble performance",
        "Has no effect"
      ],
      "correctAnswerIndex": 2,
      "explanation": "More diverse models provide uncorrelated errors, which improves averaging and ensemble performance."
    },
    {
      "id": 62,
      "questionText": "Bagging is considered a parallel ensemble method because:",
      "options": [
        "Feature selection is sequential",
        "Bootstrap samples are dependent",
        "All models are trained independently",
        "Models are trained sequentially"
      ],
      "correctAnswerIndex": 2,
      "explanation": "In Bagging, models are trained independently on different bootstrap samples, allowing parallel computation."
    },
    {
      "id": 63,
      "questionText": "Bagging performance is limited by:",
      "options": [
        "Dataset size",
        "Correlation among base models",
        "Feature normalization",
        "Bias of base models only"
      ],
      "correctAnswerIndex": 1,
      "explanation": "If base models are highly correlated, averaging them does not reduce variance effectively, limiting Bagging performance."
    },
    {
      "id": 64,
      "questionText": "When would increasing bootstrap sample size improve Bagging?",
      "options": [
        "When bias is too low",
        "When individual models are undertrained",
        "When model is overfitting",
        "When using boosting"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Larger bootstrap samples provide better training for each base model, improving overall ensemble performance."
    },
    {
      "id": 65,
      "questionText": "Which scenario reduces Bagging effectiveness?",
      "options": [
        "Large datasets",
        "High variance models",
        "Deep decision trees",
        "Highly correlated base models"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Highly correlated base models reduce the benefit of averaging predictions, making Bagging less effective."
    },
    {
      "id": 66,
      "questionText": "Bagging can be implemented for regression using:",
      "options": [
        "PCA only",
        "Only linear regression",
        "Decision trees or other regressors",
        "Clustering algorithms"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging can be applied with any high variance regressor, commonly decision trees."
    },
    {
      "id": 67,
      "questionText": "How does Bagging affect model variance?",
      "options": [
        "Leaves variance unchanged",
        "Increases variance",
        "Reduces variance",
        "Reduces bias only"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Averaging predictions from multiple models reduces variance compared to individual base models."
    },
    {
      "id": 68,
      "questionText": "Which is true about Bagging and Random Forest?",
      "options": [
        "Random Forest increases bias",
        "Random Forest is sequential boosting",
        "Random Forest is Bagging with feature randomness",
        "Random Forest has no bootstrap"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Random Forest builds on Bagging and adds random feature selection to reduce tree correlation."
    },
    {
      "id": 69,
      "questionText": "What type of learners are less likely to benefit from Bagging?",
      "options": [
        "Stable, low-variance learners",
        "Deep learners",
        "High-variance trees",
        "Noisy models"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Stable, low-variance models already produce consistent predictions; Bagging adds little benefit."
    },
    {
      "id": 70,
      "questionText": "Which factor does NOT influence Bagging effectiveness?",
      "options": [
        "Correlation among models",
        "Feature scaling",
        "Diversity of models",
        "Number of base models"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Bagging effectiveness is influenced by model diversity, correlation, and number of models, but feature scaling does not play a direct role."
    },
    {
      "id": 71,
      "questionText": "You have a high-dimensional dataset with correlated features. How would Bagging performance be affected?",
      "options": [
        "Performance is unaffected",
        "Bias will reduce significantly",
        "Performance may degrade due to correlation among base models",
        "Performance will improve automatically"
      ],
      "correctAnswerIndex": 2,
      "explanation": "High correlation among base models reduces the benefit of averaging, which can degrade Bagging performance. Random feature selection can help mitigate this."
    },
    {
      "id": 72,
      "questionText": "In a dataset with severe class imbalance, how can Bagging be adapted?",
      "options": [
        "Use balanced bootstrap samples or weighted voting",
        "Reduce number of trees",
        "Apply PCA before Bagging",
        "Ignore imbalance as Bagging handles it automatically"
      ],
      "correctAnswerIndex": 0,
      "explanation": "For imbalanced datasets, Bagging can use balanced bootstrap samples or weight the voting process to handle minority classes more effectively."
    },
    {
      "id": 73,
      "questionText": "If Bagging is applied to already overfitted deep trees, what is the likely outcome?",
      "options": [
        "Variance decreases, but predictions may still overfit slightly",
        "Overfitting increases",
        "Bias decreases significantly",
        "Model becomes linear"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Bagging reduces variance of overfitted models, stabilizing predictions, but extreme overfitting may still persist to some extent."
    },
    {
      "id": 74,
      "questionText": "Which is a real-world scenario where Bagging might fail?",
      "options": [
        "High variance decision trees",
        "Noisy datasets",
        "Small datasets with low variance models",
        "Large datasets"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging relies on diversity from bootstrap samples. Small datasets with low variance models do not benefit much, limiting Bagging effectiveness."
    },
    {
      "id": 75,
      "questionText": "How does Bagging compare to boosting in terms of error reduction?",
      "options": [
        "Both reduce variance only",
        "Bagging reduces variance, boosting reduces bias",
        "Bagging reduces bias, boosting reduces variance",
        "Both reduce bias only"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Bagging is designed to reduce variance by averaging predictions, while boosting sequentially reduces bias by focusing on errors."
    },
    {
      "id": 76,
      "questionText": "In a scenario where computation is limited, what trade-off exists for Bagging?",
      "options": [
        "Bias increases automatically",
        "Fewer base models reduce computation but may increase variance",
        "More base models reduce computation",
        "Bootstrap sampling becomes unnecessary"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Reducing the number of models saves computation but decreases variance reduction, which may affect performance."
    },
    {
      "id": 77,
      "questionText": "Bagging is applied to a time-series prediction problem. What caution should be taken?",
      "options": [
        "Bootstrap samples should respect temporal order",
        "Features should be normalized first",
        "Time-series data does not need Bagging",
        "Standard bootstrap is sufficient"
      ],
      "correctAnswerIndex": 0,
      "explanation": "In time-series data, random bootstrap may break temporal relationships. Resampling should maintain temporal order."
    },
    {
      "id": 78,
      "questionText": "When using Bagging with regression trees, which is true about overfitting?",
      "options": [
        "Bagging increases overfitting",
        "Overfitting is only reduced if dataset is small",
        "Bagging has no effect on overfitting",
        "Bagging reduces overfitting due to variance averaging"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Bagging averages predictions from multiple high-variance trees, reducing overfitting by stabilizing the output."
    },
    {
      "id": 79,
      "questionText": "A Bagging model shows poor performance on unseen data. Which is the likely reason?",
      "options": [
        "Base models are biased or low variance",
        "Random feature selection is used",
        "Bootstrap sampling is perfect",
        "Number of trees is too high"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Bagging is effective for high variance models. If base models are biased or too simple, Bagging cannot improve generalization much."
    },
    {
      "id": 80,
      "questionText": "Which scenario demonstrates Bagging’s strength?",
      "options": [
        "PCA datasets",
        "Clustering datasets",
        "Small, linear datasets",
        "High variance, non-linear datasets"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Bagging excels with high variance, complex datasets, like non-linear relationships captured by decision trees."
    },
    {
      "id": 81,
      "questionText": "In a real-time prediction system, what is a potential drawback of Bagging?",
      "options": [
        "Prediction latency due to multiple models",
        "Bias increases significantly",
        "Randomness is removed",
        "Bootstrap sampling fails"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Bagging requires predictions from multiple models, which can increase latency in real-time applications."
    },
    {
      "id": 82,
      "questionText": "How can Bagging be optimized for large-scale datasets?",
      "options": [
        "Use a single base model",
        "Avoid bootstrap sampling",
        "Reduce the number of features",
        "Parallelize model training across processors"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Bagging can be parallelized because each model is trained independently, making it scalable for large datasets."
    },
    {
      "id": 83,
      "questionText": "If base models are highly correlated, which approach can improve Bagging?",
      "options": [
        "Reduce tree depth",
        "Use single model",
        "Random feature selection (like Random Forest)",
        "Increase sample size only"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Introducing feature randomness decreases correlation among models, improving Bagging’s effectiveness."
    },
    {
      "id": 84,
      "questionText": "Bagging is applied to image classification with deep trees. Which is a valid advantage?",
      "options": [
        "Reduces variance while capturing complex patterns",
        "Decreases number of features",
        "Removes need for normalization",
        "Reduces dataset size"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Bagging stabilizes predictions from complex trees while still allowing each tree to capture intricate patterns."
    },
    {
      "id": 85,
      "questionText": "Which of the following scenarios benefits least from Bagging?",
      "options": [
        "High variance decision trees",
        "Noisy data with high variance trees",
        "Low variance models like linear regression",
        "Classification tasks with deep trees"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Stable, low-variance models do not gain significant improvement from Bagging, as variance is already low."
    },
    {
      "id": 86,
      "questionText": "How does Bagging handle overfitting in ensemble models?",
      "options": [
        "Ignores overfitting completely",
        "Increases it by adding more models",
        "Reduces it by averaging multiple high variance models",
        "Reduces bias instead of variance"
      ],
      "correctAnswerIndex": 2,
      "explanation": "By averaging predictions from multiple overfitted models, Bagging reduces variance and helps mitigate overfitting."
    },
    {
      "id": 87,
      "questionText": "What is the main difference between Random Forest and standard Bagging?",
      "options": [
        "Random Forest uses boosting instead",
        "Random Forest has no bootstrap samples",
        "Random Forest adds feature randomness at splits",
        "Random Forest reduces bias instead of variance"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Random Forest builds upon Bagging by introducing random feature selection at each split to reduce correlation among trees."
    },
    {
      "id": 88,
      "questionText": "When Bagging is used with regression trees on large noisy datasets, what is the effect?",
      "options": [
        "Training time decreases",
        "Variance is reduced, predictions are more stable",
        "Models always overfit",
        "Bias is eliminated completely"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Bagging averages predictions from multiple trees, reducing variance and stabilizing outputs even in noisy datasets."
    },
    {
      "id": 89,
      "questionText": "In practice, what is a reason to limit the number of trees in Bagging?",
      "options": [
        "Computational cost and diminishing returns on variance reduction",
        "Randomness is lost",
        "Bias increases automatically",
        "Training becomes sequential"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Beyond a certain point, adding more trees does not significantly reduce variance but increases computation."
    },
    {
      "id": 90,
      "questionText": "In which scenario is Bagging most likely to fail?",
      "options": [
        "High-variance decision trees",
        "Large-scale datasets with parallel computation",
        "Low-variance, biased base learners",
        "Noisy datasets"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Bagging reduces variance; it cannot fix high-bias, low-variance models, which limits its effectiveness."
    },
    {
      "id": 91,
      "questionText": "You want to reduce prediction variance for a stock market model using trees. What method should you consider?",
      "options": [
        "Clustering",
        "PCA only",
        "Single linear regression",
        "Bagging ensemble of decision trees"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Stock market predictions are high-variance. Bagging multiple decision trees stabilizes predictions and reduces variance."
    },
    {
      "id": 92,
      "questionText": "For highly correlated features, which Bagging modification helps performance?",
      "options": [
        "Remove bootstrap",
        "Random feature selection at splits",
        "Use shallow trees",
        "Increase number of estimators only"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Random feature selection reduces correlation among trees, improving the effectiveness of Bagging."
    },
    {
      "id": 93,
      "questionText": "Which is a computational challenge with Bagging?",
      "options": [
        "Bias increases automatically",
        "Training multiple models increases time and memory",
        "Overfitting is unavoidable",
        "Bootstrap sampling fails on large datasets"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Training many models independently can be computationally intensive, especially for large datasets."
    },
    {
      "id": 94,
      "questionText": "In a classification problem with Bagging, why might majority voting fail?",
      "options": [
        "If features are normalized",
        "If base models are biased or misclassify the same instances",
        "If dataset is large",
        "If number of trees is too high"
      ],
      "correctAnswerIndex": 1,
      "explanation": "If all base models are biased in the same way, majority voting will not correct the errors."
    },
    {
      "id": 95,
      "questionText": "Bagging is considered robust because:",
      "options": [
        "Outliers have reduced impact due to averaging",
        "Bootstrap samples are ignored",
        "Correlation is increased",
        "Bias is eliminated"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Averaging predictions reduces the effect of outliers, making Bagging more robust to noisy data."
    },
    {
      "id": 96,
      "questionText": "Which scenario illustrates Bagging’s limitation?",
      "options": [
        "Using stable low-variance models where averaging provides minimal gain",
        "Using high variance models",
        "Using noisy datasets",
        "Using parallel computation"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Stable low-variance models do not benefit from Bagging as variance is already low."
    },
    {
      "id": 97,
      "questionText": "In Bagging, if base models perform differently on subsets of data, what is the effect?",
      "options": [
        "Prediction variance decreases and ensemble is more stable",
        "Training time reduces",
        "Ensemble fails completely",
        "Bias increases dramatically"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Diverse base models provide uncorrelated errors; averaging reduces variance and stabilizes predictions."
    },
    {
      "id": 98,
      "questionText": "How can Bagging handle noisy labels in training data?",
      "options": [
        "Models ignore noisy samples",
        "Noise is amplified automatically",
        "Bias is eliminated completely",
        "Averaging predictions reduces the impact of noise"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Averaging predictions from multiple models mitigates the effect of noisy labels in the final output."
    },
    {
      "id": 99,
      "questionText": "Which factor can limit Bagging effectiveness in real-world applications?",
      "options": [
        "Bootstrap sampling",
        "High correlation among base learners",
        "High variance of base learners",
        "Parallel training"
      ],
      "correctAnswerIndex": 1,
      "explanation": "High correlation among base models reduces variance reduction, limiting Bagging performance."
    },
    {
      "id": 100,
      "questionText": "Which is a key consideration when applying Bagging to real-world regression problems?",
      "options": [
        "Bagging always guarantees perfect predictions",
        "Only number of features matters",
        "Base model complexity, number of estimators, and correlation among models",
        "Bootstrap size is irrelevant"
      ],
      "correctAnswerIndex": 2,
      "explanation": "For effective Bagging, you must consider base model complexity, ensemble size, and model correlation to ensure variance reduction and generalization."
    }
  ]
}