data/Voting.json

{
  "title": "Voting Ensemble Mastery: 100 MCQs",
  "description": "A complete MCQ set on Voting Ensemble Methods — covering hard and soft voting, use-cases, advantages, limitations, and real-world scenario questions.",
  "questions": [
    {
      "id": 1,
      "questionText": "What is the core idea of a Voting Ensemble?",
      "options": [
        "Train a single strong model",
        "Perform dimensionality reduction",
        "Reduce dataset size",
        "Combine predictions from multiple models"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Voting Ensembling combines predictions from multiple models to improve overall accuracy."
    },
    {
      "id": 2,
      "questionText": "What are the two main types of Voting?",
      "options": [
        "Bagging and Boosting",
        "Static and Dynamic Voting",
        "Linear and Non-linear Voting",
        "Hard and Soft Voting"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Voting Ensembles are primarily divided into Hard Voting and Soft Voting methods."
    },
    {
      "id": 3,
      "questionText": "What does Hard Voting use to make the final prediction?",
      "options": [
        "Highest loss",
        "Majority class vote",
        "Average probabilities",
        "Gradient values"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Hard Voting chooses the class that appears most frequently among model predictions."
    },
    {
      "id": 4,
      "questionText": "Soft Voting makes predictions based on:",
      "options": [
        "Averaging class probabilities",
        "Majority class votes",
        "Random selection",
        "Model with highest accuracy only"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Soft Voting averages probabilities from all models and selects the class with the highest probability."
    },
    {
      "id": 5,
      "questionText": "Soft Voting requires that base models must:",
      "options": [
        "Have the same accuracy",
        "Output raw labels",
        "Output probability scores",
        "Be decision trees only"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Soft Voting needs probability outputs like `predict_proba()` — not just class labels."
    },
    {
      "id": 6,
      "questionText": "What is the minimum number of models required for a Voting Ensemble?",
      "options": [
        "3",
        "1",
        "2",
        "No minimum"
      ],
      "correctAnswerIndex": 2,
      "explanation": "At least 2 models are required to perform any kind of voting."
    },
    {
      "id": 7,
      "questionText": "What is the purpose of using multiple models in Voting?",
      "options": [
        "To combine strengths of different models",
        "To reduce dataset size",
        "To increase bias",
        "To make training faster"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Voting combines multiple models to leverage their strengths and improve prediction reliability."
    },
    {
      "id": 8,
      "questionText": "In Hard Voting, what happens if there is a tie between class predictions?",
      "options": [
        "First class is selected",
        "Depends on implementation",
        "Model with highest accuracy is selected",
        "Random class is selected"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Tie handling is implementation-dependent and varies across libraries."
    },
    {
      "id": 9,
      "questionText": "Which Voting method performs better when base models are calibrated and output probabilities?",
      "options": [
        "Soft Voting",
        "Hard Voting",
        "Rule-based Voting",
        "Random Voting"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Soft Voting uses probability averaging — works best with well-calibrated models."
    },
    {
      "id": 10,
      "questionText": "Which of the following is a key advantage of Voting over a single model?",
      "options": [
        "Requires less computation",
        "Better generalization",
        "Always 100% accuracy",
        "No need for tuning"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Voting reduces the chance of overfitting and improves generalization performance."
    },
    {
      "id": 11,
      "questionText": "Which type of Voting is preferred when class probabilities are reliable?",
      "options": [
        "Bootstrap Voting",
        "Soft Voting",
        "Random Voting",
        "Hard Voting"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Soft Voting utilizes probability outputs effectively when they are calibrated."
    },
    {
      "id": 12,
      "questionText": "What is a requirement for models in Soft Voting?",
      "options": [
        "All models must be trees",
        "All models must be neural networks",
        "All models must use same hyperparameters",
        "All models must output probability scores"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Soft Voting needs probability scores using functions like `predict_proba()`."
    },
    {
      "id": 13,
      "questionText": "Which is true about Voting Ensembles?",
      "options": [
        "They are only used for regression",
        "They must use only identical models",
        "They reduce overfitting by aggregating independent models",
        "They eliminate the need for feature engineering"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Voting reduces overfitting by combining diverse independent models."
    },
    {
      "id": 14,
      "questionText": "What type of models can be used inside a Voting Ensemble?",
      "options": [
        "Only decision trees",
        "Only SVM",
        "Any mix of models (heterogeneous)",
        "Only neural networks"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Voting can combine diverse models like SVM, Logistic Regression, Decision Trees, etc."
    },
    {
      "id": 15,
      "questionText": "Which problem type is Voting Ensemble typically used for?",
      "options": [
        "Only regression",
        "Only classification",
        "Both classification and regression",
        "Only clustering"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Voting is mainly used for classification, but can also be extended to regression."
    },
    {
      "id": 16,
      "questionText": "In Hard Voting, how is the final class decided?",
      "options": [
        "By averaging probabilities",
        "By selecting random model output",
        "By selecting highest confidence model",
        "By selecting majority voted class labels"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Hard Voting simply picks the class label that gets majority votes."
    },
    {
      "id": 17,
      "questionText": "Which of the following is a limitation of Voting?",
      "options": [
        "Always requires GPUs",
        "Not interpretable easily",
        "Can be used only with CNNs",
        "Cannot handle classification tasks"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Since multiple models are used, analyzing why a prediction was made becomes harder."
    },
    {
      "id": 18,
      "questionText": "Which is true for Hard Voting?",
      "options": [
        "Requires all models to be identical",
        "Needs only class labels",
        "Slower than Soft Voting",
        "Uses probabilities"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Hard Voting only needs class labels like 0/1, not probabilities."
    },
    {
      "id": 19,
      "questionText": "Is it possible to combine Logistic Regression, SVM, and Random Forest in a Voting Ensemble?",
      "options": [
        "Only if dataset is small",
        "Yes, heterogeneous models are allowed",
        "Only if all are deep learning models",
        "No, all models must be same type"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Voting allows combining different model families for better performance."
    },
    {
      "id": 20,
      "questionText": "What does Soft Voting average?",
      "options": [
        "Model parameters",
        "Raw model inputs",
        "Dataset rows",
        "Predicted class probabilities"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Soft Voting averages probability outputs, then selects highest probability class."
    },
    {
      "id": 21,
      "questionText": "Which library provides VotingClassifier in Python?",
      "options": [
        "NumPy",
        "PyTorch",
        "scikit-learn",
        "TensorFlow"
      ],
      "correctAnswerIndex": 2,
      "explanation": "scikit-learn provides VotingClassifier for ensembling models."
    },
    {
      "id": 22,
      "questionText": "Which voting type is more robust against noisy class probability estimations?",
      "options": [
        "Hard Voting",
        "Soft Voting",
        "None",
        "Random Voting"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Hard Voting is safer when probabilities are unreliable or poorly calibrated."
    },
    {
      "id": 23,
      "questionText": "Can Voting Ensembles improve stability of model predictions?",
      "options": [
        "Only for time series",
        "Yes, by reducing variance",
        "Only in regression",
        "No, increases randomness"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Voting helps reduce variance by averaging multiple predictions."
    },
    {
      "id": 24,
      "questionText": "If models strongly disagree in Hard Voting, what happens?",
      "options": [
        "Soft Voting is automatically used",
        "Prediction becomes unstable",
        "Voting skips such cases",
        "It stops training"
      ],
      "correctAnswerIndex": 1,
      "explanation": "High disagreement can reduce prediction confidence and stability."
    },
    {
      "id": 25,
      "questionText": "What happens if one weak model is added to Soft Voting?",
      "options": [
        "No effect at all",
        "Causes overfitting immediately",
        "Always improves accuracy",
        "Can reduce overall performance"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Soft Voting averages probabilities — a weak noisy model can hurt accuracy."
    },
    {
      "id": 26,
      "questionText": "What is the main difference between Bagging and Voting?",
      "options": [
        "Bagging is only for regression",
        "Voting always boosts performance, Bagging does not",
        "Voting uses different model types, Bagging uses same model type",
        "Voting needs large datasets only"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Voting is heterogeneous by nature; Bagging generally uses the same model type with data bootstrapping."
    },
    {
      "id": 27,
      "questionText": "Which statement is true about Hard Voting?",
      "options": [
        "It uses average probability",
        "It trains models sequentially",
        "It requires all models to be deep learning models",
        "It selects the majority class label"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Hard voting picks the class label that appears most frequently among model predictions."
    },
    {
      "id": 28,
      "questionText": "Soft Voting is more reliable than Hard Voting when:",
      "options": [
        "The dataset is extremely small",
        "Model probabilities are well-calibrated",
        "Using only one model",
        "Class labels are noisy"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Soft voting performs best when model probability outputs are accurate."
    },
    {
      "id": 29,
      "questionText": "Which of the following is a key advantage of Soft Voting?",
      "options": [
        "Avoids probability calculations entirely",
        "Can weigh models differently using probabilities",
        "Only works with random forests",
        "Does not need probability estimates"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Soft voting can assign more importance to stronger models using weighted averaging."
    },
    {
      "id": 30,
      "questionText": "Voting Ensemble works best when base models are:",
      "options": [
        "From the same algorithm",
        "Poorly trained",
        "Highly correlated",
        "Diverse and independent"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Model diversity ensures different error patterns, improving ensemble performance."
    },
    {
      "id": 31,
      "questionText": "In Voting Ensemble, combining Logistic Regression, SVM, and Decision Tree is an example of:",
      "options": [
        "Bagging",
        "Sequential ensemble",
        "Heterogeneous ensemble",
        "Homogeneous ensemble"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Using different types of models is called a heterogeneous ensemble."
    },
    {
      "id": 32,
      "questionText": "Which type of Voting allows assigning more importance to better-performing models?",
      "options": [
        "Uniform Voting",
        "Random Voting",
        "Hard Voting",
        "Soft Voting with weights"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Soft voting supports weighting individual models based on performance."
    },
    {
      "id": 33,
      "questionText": "What happens if one model in a Voting Ensemble consistently gives wrong predictions?",
      "options": [
        "It fully controls the final output",
        "It stops the ensemble from working",
        "It improves accuracy",
        "It slightly decreases overall accuracy"
      ],
      "correctAnswerIndex": 3,
      "explanation": "A weak model can slightly hurt performance but often the ensemble still performs well."
    },
    {
      "id": 34,
      "questionText": "Which Voting method is more interpretable regarding final decision logic?",
      "options": [
        "Both equally",
        "None",
        "Hard Voting",
        "Soft Voting"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Hard voting decisions can be directly traced to majority class votes."
    },
    {
      "id": 35,
      "questionText": "Soft Voting may underperform if:",
      "options": [
        "Models are shallow",
        "All models agree",
        "Models don't output probabilities",
        "Dataset is small"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Soft voting requires probability outputs like predict_proba()."
    },
    {
      "id": 36,
      "questionText": "Which of these is a REAL requirement for Soft Voting?",
      "options": [
        "All models must be trees",
        "All models must be neural networks",
        "All models must output class probabilities",
        "All models must have same accuracy"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Soft voting requires probability estimates — no need for same accuracy or model type."
    },
    {
      "id": 37,
      "questionText": "Voting Ensembles improve performance mainly by reducing:",
      "options": [
        "Bias",
        "Training time",
        "Variance",
        "Dataset size"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Voting helps reduce the variance of predictions by averaging model outputs."
    },
    {
      "id": 38,
      "questionText": "Which is a potential DISADVANTAGE of Voting Ensembles?",
      "options": [
        "Cannot be used for classification",
        "Hard to interpret final decisions",
        "Must use same model type",
        "Only works for regression"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Since multiple models influence the result, interpretability decreases."
    },
    {
      "id": 39,
      "questionText": "Voting Ensemble is MOST helpful when individual models:",
      "options": [
        "Have identical predictions",
        "Use the same algorithm and parameters",
        "Are all overfitted",
        "Make complementary errors"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Ensemble works best when models compensate for each other's errors."
    },
    {
      "id": 40,
      "questionText": "Which scenario best fits using Voting Ensemble?",
      "options": [
        "Multiple trained models with decent accuracy",
        "Highly imbalanced dataset with no labels",
        "Real-time system with tight latency constraint",
        "Only one extremely accurate model"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Voting is useful when several decent but imperfect models are available."
    },
    {
      "id": 41,
      "questionText": "Which of the following is TRUE about weighting in Soft Voting?",
      "options": [
        "Weights decrease ensemble accuracy always",
        "Weights are randomly assigned",
        "Weights are only used in Hard Voting",
        "Weights allow stronger models to influence more"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Weights in Soft Voting help give preference to stronger models."
    },
    {
      "id": 42,
      "questionText": "Hard Voting is most effective when:",
      "options": [
        "Models produce random outputs",
        "Dataset is unsupervised",
        "Class labels from models are stable",
        "Probabilities are reliable"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Hard voting is useful when class labels are confidently predicted."
    },
    {
      "id": 43,
      "questionText": "Which approach improves Soft Voting performance?",
      "options": [
        "Avoid probability averaging",
        "Use uncalibrated probability models",
        "Use calibrated probability models",
        "Remove probability outputs"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Soft voting needs properly calibrated probability outputs for reliability."
    },
    {
      "id": 44,
      "questionText": "In voting, which models are preferred for maximum accuracy gain?",
      "options": [
        "Strong but diverse models",
        "Extremely similar models",
        "Very weak models only",
        "Highly correlated models"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Diversity ensures different error patterns, maximizing ensemble accuracy."
    },
    {
      "id": 45,
      "questionText": "What is a potential RISK of including too many models in a Voting Ensemble?",
      "options": [
        "Loss of supervised learning",
        "Higher computation and latency",
        "Automatic model deletion",
        "Overfitting on test data"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Too many models increase compute cost and response time."
    },
    {
      "id": 46,
      "questionText": "Which situation can DEGRADE Voting Ensemble performance?",
      "options": [
        "Using only calibrated probability models",
        "Adding several almost identical models",
        "Combining different feature extractors",
        "Adding multiple weak uncorrelated learners"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Redundant identical models bring no diversity and give no benefit."
    },
    {
      "id": 47,
      "questionText": "Soft Voting is preferred over Hard Voting when:",
      "options": [
        "Only labels are needed",
        "No model supports probability output",
        "Probability outputs are reliable",
        "Models are identical"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Soft voting requires accurate probability outputs for better performance."
    },
    {
      "id": 48,
      "questionText": "If accuracy of individual models is low but different mistakes are made, Voting can still:",
      "options": [
        "Stop working",
        "Always fail",
        "Outperform individual models",
        "Perform worse than all models"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Ensemble effect combines different strengths, even if individual accuracy is modest."
    },
    {
      "id": 49,
      "questionText": "Which type of dataset benefits MOST from Voting Ensemble?",
      "options": [
        "Large and diverse structured data",
        "Purely unstructured images only",
        "Datasets with no labels",
        "Dataset with only one feature"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Voting is very effective in structured tabular data problems."
    },
    {
      "id": 50,
      "questionText": "Which metric is NOT directly improved by Voting Ensemble?",
      "options": [
        "Robustness",
        "Stability",
        "Training speed",
        "Generalization"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Voting increases computation; it does not accelerate training."
    },
    {
      "id": 51,
      "questionText": "Medium-Level: Soft Voting gives better performance over Hard Voting when:",
      "options": [
        "Model diversity does not exist",
        "Ensemble contains a single model",
        "Model probabilities are well calibrated",
        "Models only provide class labels"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Soft voting leverages probability information, so accurate calibrated probabilities improve performance."
    },
    {
      "id": 52,
      "questionText": "Which strategy improves Voting Ensemble performance?",
      "options": [
        "Using only identical models",
        "Skipping data preprocessing",
        "Blending diverse model architectures",
        "Ignoring validation scores"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Diversity among models boosts ensemble power significantly."
    },
    {
      "id": 53,
      "questionText": "Voting Ensemble can fail if base models are:",
      "options": [
        "Moderately accurate",
        "Trained on different features",
        "Diverse and independent",
        "Highly correlated with similar errors"
      ],
      "correctAnswerIndex": 3,
      "explanation": "If base models make similar errors, voting does not reduce errors effectively."
    },
    {
      "id": 54,
      "questionText": "Soft Voting with weights allows:",
      "options": [
        "Random selection of predictions",
        "Ignore weaker models completely",
        "Greater influence for stronger models",
        "Equal influence for all models"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Weights in Soft Voting give more importance to better-performing models."
    },
    {
      "id": 55,
      "questionText": "Hard Voting is more robust when:",
      "options": [
        "The dataset is very large",
        "Individual model predictions are noisy",
        "There is only one base model",
        "Model probabilities are perfect"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Hard Voting reduces sensitivity to probability estimation errors by using majority votes."
    },
    {
      "id": 56,
      "questionText": "Which scenario is best suited for a Voting Ensemble?",
      "options": [
        "Several moderately performing models with different strengths exist",
        "All models are identical",
        "Only one high-performing model is available",
        "Dataset has no labels"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Voting leverages complementary predictions from multiple models to improve accuracy."
    },
    {
      "id": 57,
      "questionText": "When combining models in Voting, diversity helps to:",
      "options": [
        "Decrease overall variance",
        "Reduce dataset size",
        "Increase bias",
        "Accelerate training"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Diverse models make different errors, which are averaged out, reducing variance."
    },
    {
      "id": 58,
      "questionText": "Adding very weak models to a Voting Ensemble can:",
      "options": [
        "Have no effect",
        "Always improve accuracy",
        "Slightly reduce overall performance",
        "Break the ensemble"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Weak models may introduce noise and slightly lower ensemble accuracy."
    },
    {
      "id": 59,
      "questionText": "In a Voting Ensemble, tie-breaking in Hard Voting depends on:",
      "options": [
        "Number of features",
        "Probability outputs",
        "Implementation details",
        "Dataset size"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Hard Voting tie-handling is usually implementation-specific."
    },
    {
      "id": 60,
      "questionText": "Voting Ensemble helps improve:",
      "options": [
        "Underfitting only",
        "Training speed",
        "Generalization and robustness",
        "Overfitting only"
      ],
      "correctAnswerIndex": 2,
      "explanation": "By combining multiple models, ensembles improve generalization and reduce sensitivity to noise."
    },
    {
      "id": 61,
      "questionText": "Which base models can be combined in a Voting Ensemble?",
      "options": [
        "Any heterogeneous or homogeneous models",
        "Only neural networks",
        "Only decision trees",
        "Only linear models"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Voting allows combining different model types for better ensemble performance."
    },
    {
      "id": 62,
      "questionText": "Medium Level: If one base model in Soft Voting produces biased probabilities, the ensemble will:",
      "options": [
        "Ignore the model automatically",
        "Switch to Hard Voting",
        "Always fail",
        "Average out if others are unbiased"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Soft Voting averages probabilities, so other unbiased models can compensate for one biased model."
    },
    {
      "id": 63,
      "questionText": "Scenario: You have three classifiers, two strong and one weak. Soft Voting is used. Which is true?",
      "options": [
        "Soft Voting ignores weak models",
        "Strong models have greater influence if weighted",
        "All models have equal effect regardless of performance",
        "Weak model dominates ensemble"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Weighted Soft Voting allows strong models to have higher influence on the final prediction."
    },
    {
      "id": 64,
      "questionText": "Scenario: A Voting Ensemble has low diversity. Expected outcome?",
      "options": [
        "Significant increase in accuracy",
        "Low improvement over individual models",
        "High variance reduction",
        "Ensemble stops working"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Without diversity, models make similar errors and the ensemble gains little benefit."
    },
    {
      "id": 65,
      "questionText": "Medium Level: How to handle class imbalance in Voting Ensemble?",
      "options": [
        "Ignore imbalance",
        "Remove minority class",
        "Use class weighting or balanced sampling in base models",
        "Only use Hard Voting"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Adjusting base models to handle class imbalance ensures the ensemble is not biased."
    },
    {
      "id": 66,
      "questionText": "Scenario: Soft Voting probabilities differ in scale among models. Recommended step?",
      "options": [
        "Remove some models",
        "Use only Hard Voting",
        "Normalize probabilities before averaging",
        "Ignore scaling differences"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Scaling ensures probabilities are comparable before combining in Soft Voting."
    },
    {
      "id": 67,
      "questionText": "Scenario: Voting Ensemble shows unstable predictions on edge cases. Likely reason?",
      "options": [
        "Using Hard Voting",
        "Training dataset too large",
        "Too many base models",
        "Insufficient diversity in base models"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Lack of diversity leads to correlated errors, causing instability."
    },
    {
      "id": 68,
      "questionText": "Scenario: You want a lightweight Voting model for real-time use. Best practice?",
      "options": [
        "Ignore computation constraints",
        "Add more complex models",
        "Reduce number of base models and simplify them",
        "Use Soft Voting only"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Fewer and simpler models reduce latency while maintaining reasonable ensemble performance."
    },
    {
      "id": 69,
      "questionText": "Scenario: Hard Voting ensemble has many ties. Recommended action?",
      "options": [
        "Remove base models",
        "Switch to Soft Voting or assign weights",
        "Keep Hard Voting as is",
        "Randomly select predictions"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Soft Voting or weighting reduces the effect of ties and improves reliability."
    },
    {
      "id": 70,
      "questionText": "Scenario: You combine Logistic Regression, Random Forest, and SVM using Soft Voting. Test accuracy is lower than best base model. Possible causes?",
      "options": [
        "Soft Voting always underperforms",
        "Random initialization of models",
        "Improper probability calibration or correlated errors",
        "Training data too large"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Soft Voting performance depends on probability calibration and diversity among base models."
    },
    {
      "id": 71,
      "questionText": "Hard Voting vs Soft Voting: Which is better for well-calibrated probabilistic outputs?",
      "options": [
        "Neither",
        "Both equal",
        "Soft Voting",
        "Hard Voting"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Soft Voting leverages probability outputs and generally performs better with well-calibrated models."
    },
    {
      "id": 72,
      "questionText": "Scenario: One base model in a Voting Ensemble fails completely on new data. Ensemble effect?",
      "options": [
        "Hard Voting stops working",
        "Soft Voting averages can reduce impact",
        "Ensemble accuracy drops to zero",
        "All models ignored"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Soft Voting can mitigate the effect of one failing model by averaging with other models' outputs."
    },
    {
      "id": 73,
      "questionText": "Scenario: Ensemble predictions fluctuate across runs. Likely cause?",
      "options": [
        "Multiple base models",
        "Random initialization of base models",
        "High dataset diversity",
        "Using Soft Voting"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Randomness in training some base models can cause prediction fluctuations."
    },
    {
      "id": 74,
      "questionText": "Scenario: Weighted Soft Voting used incorrectly. What could happen?",
      "options": [
        "Hard Voting automatically applies",
        "Strong models underrepresented, ensemble underperforms",
        "Ensemble accuracy always increases",
        "Base models ignored"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Incorrect weighting can reduce the influence of stronger models, decreasing ensemble performance."
    },
    {
      "id": 75,
      "questionText": "Scenario: Using Soft Voting with outputs on different scales. What to do?",
      "options": [
        "Randomly select one model",
        "Use Hard Voting",
        "Normalize probabilities",
        "Ignore scaling"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Normalizing probabilities ensures a fair contribution from all models."
    },
    {
      "id": 76,
      "questionText": "Scenario: Ensemble overfits on training data. Recommended solution?",
      "options": [
        "Ignore ensemble and use single model",
        "Use cross-validation and consider reducing base models or regularizing them",
        "Switch to Hard Voting only",
        "Add more weak models"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Proper cross-validation and controlling base model complexity help prevent overfitting."
    },
    {
      "id": 77,
      "questionText": "Scenario: Voting Ensemble performs worse than individual models. Likely reason?",
      "options": [
        "Voting always underperforms",
        "Dataset too large",
        "Using Soft Voting automatically fails",
        "Base models highly correlated or weak"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Ensemble benefits only arise when base models are diverse and reasonably accurate."
    },
    {
      "id": 78,
      "questionText": "Scenario: Ensemble used for imbalanced classification. Which strategy helps?",
      "options": [
        "Ignore class imbalance",
        "Class weighting or balanced sampling in base models",
        "Remove minority classes",
        "Use Hard Voting only"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Adjusting base models for imbalance ensures ensemble predictions are not biased."
    },
    {
      "id": 79,
      "questionText": "Scenario: Adding highly similar base models. Ensemble outcome?",
      "options": [
        "Maximum accuracy gain",
        "Soft Voting fails",
        "Little to no improvement",
        "Hard Voting fails"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Similar models add redundancy, giving minimal benefit to the ensemble."
    },
    {
      "id": 80,
      "questionText": "Scenario: Ensemble shows different accuracy across runs. Most likely reason?",
      "options": [
        "Hard Voting always fluctuates",
        "Soft Voting inherently unstable",
        "Randomness in training base models",
        "Dataset size too small"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Random initialization or stochastic training of base models causes variance across runs."
    },
    {
      "id": 81,
      "questionText": "Scenario: You need an ensemble for high-risk decisions where errors are costly. Best approach?",
      "options": [
        "Random Voting",
        "Hard Voting with weak models",
        "Weighted Soft Voting with strong base models",
        "Single uncalibrated model"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Weighted Soft Voting emphasizes reliable models and reduces error in critical decisions."
    },
    {
      "id": 82,
      "questionText": "Scenario: You combine models trained on overlapping features. Risk?",
      "options": [
        "Soft Voting ignored",
        "Ensemble fails to produce output",
        "Highly correlated errors, reduced ensemble benefit",
        "Maximum ensemble improvement"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Correlated errors reduce the variance reduction benefit of the ensemble."
    },
    {
      "id": 83,
      "questionText": "Scenario: One model outputs extreme probabilities. Soft Voting effect?",
      "options": [
        "Hard Voting preferred",
        "May skew average unless weights or normalization applied",
        "Has no effect",
        "Automatically corrected"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Extreme probabilities can dominate averaging; normalization or weights correct this."
    },
    {
      "id": 84,
      "questionText": "Scenario: Voting Ensemble uses both linear and nonlinear models. Expected benefit?",
      "options": [
        "Soft Voting ignored",
        "No benefit, models cancel each other",
        "Capture complex patterns better than single model type",
        "Ensemble fails"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Heterogeneous ensembles leverage strengths of diverse models."
    },
    {
      "id": 85,
      "questionText": "Scenario: You observe overfitting in ensemble predictions. Recommended step?",
      "options": [
        "Regularize base models and limit number of learners",
        "Switch from Soft to Hard Voting",
        "Ignore and keep current setup",
        "Add more weak base models"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Controlling base model complexity reduces overfitting risk in the ensemble."
    },
    {
      "id": 86,
      "questionText": "Scenario: Voting Ensemble applied on noisy data. Best choice?",
      "options": [
        "Use only one model",
        "Hard Voting may be more robust to noisy predictions",
        "Ignore noise",
        "Soft Voting always better"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Hard Voting reduces the effect of extreme probability fluctuations caused by noise."
    },
    {
      "id": 87,
      "questionText": "Scenario: You need explainability for the ensemble decision. Best choice?",
      "options": [
        "Single black-box model",
        "Hard Voting with traceable majority votes",
        "Soft Voting with unweighted averaging",
        "Use random models"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Hard Voting provides clear insight into which class won the majority vote."
    },
    {
      "id": 88,
      "questionText": "Scenario: You want to minimize latency in ensemble inference. Recommended?",
      "options": [
        "Randomize predictions",
        "Increase base models and use Soft Voting",
        "Reduce number and complexity of base models",
        "Always use Deep Neural Networks"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Fewer and simpler models decrease computation and latency while maintaining reasonable accuracy."
    },
    {
      "id": 89,
      "questionText": "Scenario: Base models trained with different feature subsets. Expected benefit?",
      "options": [
        "Reduced diversity",
        "Ensemble ignored",
        "Increased diversity and ensemble robustness",
        "Soft Voting fails"
      ],
      "correctAnswerIndex": 2,
      "explanation": "Different feature subsets create diverse predictions, improving ensemble performance."
    },
    {
      "id": 90,
      "questionText": "Scenario: You combine models with complementary strengths. Outcome?",
      "options": [
        "Enhanced performance compared to any single model",
        "Ensemble fails",
        "Performance drops",
        "Only Hard Voting works"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Combining complementary models allows the ensemble to cover weaknesses of individual models."
    },
    {
      "id": 91,
      "questionText": "Scenario: Ensemble shows slightly worse performance than best base model. Reason?",
      "options": [
        "Hard Voting ignored",
        "Models may be too correlated or weakly performing",
        "Voting always reduces performance",
        "Dataset too large"
      ],
      "correctAnswerIndex": 1,
      "explanation": "High correlation or weak base models can reduce ensemble benefit."
    },
    {
      "id": 92,
      "questionText": "Scenario: You want maximum interpretability with moderate performance. Best option?",
      "options": [
        "Random ensemble",
        "Hard Voting with simple base models",
        "Weighted Soft Voting with complex models",
        "Single complex model"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Hard Voting with simple models is easier to interpret while maintaining decent performance."
    },
    {
      "id": 93,
      "questionText": "Scenario: Ensemble prediction differs from all base models. Possible reason?",
      "options": [
        "Hard Voting tie occurs",
        "Impossible in Voting",
        "Error in data preprocessing",
        "Soft Voting probability averaging can yield a class different from all individual predictions"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Soft Voting averages probabilities, which may shift the final predicted class."
    },
    {
      "id": 94,
      "questionText": "Scenario: Using ensemble for critical medical diagnosis. Preferred setup?",
      "options": [
        "Hard Voting with weak models",
        "Single uncalibrated model",
        "Random Voting",
        "Weighted Soft Voting with calibrated models"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Weighted Soft Voting ensures reliable models dominate the prediction, improving accuracy for high-stakes tasks."
    },
    {
      "id": 95,
      "questionText": "Scenario: Ensemble uses multiple similar trees. Soft Voting vs Hard Voting?",
      "options": [
        "Soft Voting always better",
        "Hard Voting fails",
        "Ensemble ignored",
        "Little difference since models are correlated"
      ],
      "correctAnswerIndex": 3,
      "explanation": "Highly correlated models provide minimal ensemble improvement, regardless of voting type."
    },
    {
      "id": 96,
      "questionText": "Scenario: You want to balance accuracy and latency. Recommendation?",
      "options": [
        "Reduce base models and consider simpler learners",
        "Always Soft Voting",
        "Ignore latency",
        "Use all available models regardless of size"
      ],
      "correctAnswerIndex": 0,
      "explanation": "Fewer and simpler models reduce latency while maintaining reasonable accuracy."
    },
    {
      "id": 97,
      "questionText": "Scenario: Ensemble uses probabilistic outputs from calibrated models. Expected outcome?",
      "options": [
        "Soft Voting fails",
        "Improved prediction reliability using Soft Voting",
        "Hard Voting fails",
        "No difference"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Calibrated probabilities improve the effectiveness of Soft Voting."
    },
    {
      "id": 98,
      "questionText": "Scenario: Base models vary greatly in accuracy. Best Voting strategy?",
      "options": [
        "Hard Voting ignoring weights",
        "Weighted Soft Voting to emphasize stronger models",
        "Random selection",
        "Remove weaker models"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Weighted Soft Voting allows stronger models to have more influence on the final prediction."
    },
    {
      "id": 99,
      "questionText": "Scenario: Ensemble shows high variance in predictions. Possible solution?",
      "options": [
        "Reduce number of base models",
        "Increase diversity among base models",
        "Switch to single model",
        "Use uncalibrated probabilities"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Greater model diversity reduces correlated errors and stabilizes ensemble predictions."
    },
    {
      "id": 100,
      "questionText": "Scenario: You combine heterogeneous models using Voting Ensemble. Goal achieved?",
      "options": [
        "Hard Voting fails",
        "Improved generalization and robustness over individual models",
        "Soft Voting ignored",
        "Reduced accuracy"
      ],
      "correctAnswerIndex": 1,
      "explanation": "Heterogeneous ensembles leverage complementary strengths, improving generalization and robustness."
    }
  ]
}