ML Study Hub

Home / Past papers / MLP Quiz 2 / this paper

IIT M DIPLOMA AN EXAM QDD4 01 Dec 2024 - 01 Dec 2024

MLP Quiz 221 questions50 marks13 MCQ · 4 MSQ · 4 SADiploma / afternoon / QDD4
Elapsed 00:00
Q1MCQ3 marks
question figure
  1. question figure
  2. question figure
  3. question figure
  4. question figure
  5. question figure
Q2MCQ3 marks
question figure
  1. It assigns equal weightage to L1 and L2 regularizations.
  2. It applies only L2 regularization.
  3. It applies only L1 regularization.
  4. It disables regularization completely.
Q3MCQ3 marks
question figure
  1. question figure
  2. question figure
  3. question figure
  4. question figure
Q4MCQ3 marks
question figure
  1. The data will be standardized without centering, preserving the mean of thedataset.
  2. The data will be centered but not scaled, leading to variance issues.
  3. The parameter has no effect; the data will always be centered regardless ofthis setting.
  4. The scaling operation will be disabled entirely, leaving the data unchanged.
Q5MCQ2 marks
question figure
  1. question figure
  2. question figure
  3. question figure
  4. question figure
Q6MCQ2 marks
question figure
  1. question figure
  2. question figure
  3. question figure
  4. question figure
Q7MCQ2 marks
question figure
  1. 0.75
  2. 0.80
  3. 0.50
  4. 0.67
Q8MCQ2 marks
question figure
  1. LogisticRegression
  2. Perceptron
  3. RidgeClassifier
  4. Support Vector Machine (SVM)
  5. LassoClassifer
Q9MCQ2 marks
question figure
  1. question figure
  2. question figure
  3. question figure
  4. question figure
Q10MCQ2 marks
question figure
  1. The boundaries become more complex and tightly fitted to the training data.
  2. The boundaries become more regular and smooth.
  3. Decision boundaries are unaffected by C.
  4. The number of support vectors decreases.
Q11MCQ2 marks
question figure
  1. 0.99
  2. 0.87
  3. 0.95
  4. 0.80
Q12MCQ2 marks
question figure
  1. question figure
  2. question figure
  3. question figure
  4. question figure
Q13MCQ2 marks
question figure
  1. It defines the tolerance for stopping criteria.
  2. It defines the learning rate schedule.
  3. It controls regularization strength.
  4. It controls the number of iterations.
Q14MSQ2 marksselect all that apply
question figure
  1. ’linear’
  2. ’quadratic’
  3. ’rbf’
  4. ’sigmoid’
Q15MSQ2 marksselect all that apply
question figure
  1. question figure
  2. question figure
  3. question figure
  4. question figure
Q16MSQ2 marksselect all that apply
question figure
  1. question figure
  2. question figure
  3. question figure
  4. question figure
Q17MSQ3 marksselect all that apply
question figure
  1. question figure
  2. question figure
  3. question figure
  4. question figure
Q18SA3 marks
question figure
Q19SA3 marks
question figure
Q20SA3 marks
question figure
Q21SA2 marks
question figure

Answer key

  1. Q1 (MCQ, 3 marks): E
  2. Q2 (MCQ, 3 marks): A
  3. Q3 (MCQ, 3 marks): A
  4. Q4 (MCQ, 3 marks): A
  5. Q5 (MCQ, 2 marks): B
  6. Q6 (MCQ, 2 marks): B
  7. Q7 (MCQ, 2 marks): D
  8. Q8 (MCQ, 2 marks): C
  9. Q9 (MCQ, 2 marks): B
  10. Q10 (MCQ, 2 marks): A
  11. Q11 (MCQ, 2 marks): C
  12. Q12 (MCQ, 2 marks): A
  13. Q13 (MCQ, 2 marks): A
  14. Q14 (MSQ, 2 marks): A, C, D
  15. Q15 (MSQ, 2 marks): B, C, D
  16. Q16 (MSQ, 2 marks): A, C, D
  17. Q17 (MSQ, 3 marks): A, B
  18. Q18 (SA, 3 marks): 0.012 to 0.022
  19. Q19 (SA, 3 marks): 60
  20. Q20 (SA, 3 marks): 7
  21. Q21 (SA, 2 marks): 3.5

Concepts this paper tests

Auto-generated summary from the source site. Handy as a checklist; the lessons above explain each idea from scratch.

Summary of Core Topics, Concepts, Principles, and Equations

Core Topics and Concepts

  1. Machine Learning Models and Algorithms

    • Ridge Regression: A linear regression technique with L2 regularization.
    • Stochastic Gradient Descent (SGD): An optimization algorithm for training linear models.
    • Polynomial Regression: Extending linear regression by adding polynomial terms.
    • Support Vector Machines (SVM): Supervised learning models for classification and regression.
    • Logistic Regression: A classification algorithm for binary and multiclass problems.
    • K-Nearest Neighbors (KNN): A non-parametric method used for classification and regression.
    • Gaussian Naive Bayes: A probabilistic classifier based on Bayes' theorem.
    • Perceptron: A linear classifier for binary classification.
    • KNN Imputation: A method to handle missing values using KNN.

  1. Model Evaluation and Metrics

    • Mean Squared Error (MSE): A common metric for regression models.
    • R² Score: A statistical measure of how well the regression predictions approximate the real data points.
    • Precision Score: A metric for classification models, indicating the ratio of correctly predicted positive observations to the total predicted positives.

  1. Data Preprocessing and Feature Engineering

    • StandardScaler: Standardizes features by removing the mean and scaling to unit variance.
    • Polynomial Features: Generates polynomial and interaction features.
    • CountVectorizer: Converts a collection of text documents to a matrix of token counts.
    • KNN Imputation: Imputes missing values using the K-Nearest Neighbors algorithm.

  1. Hyperparameter Tuning and Model Selection

    • GridSearchCV: Exhaustive search over specified parameter values for a model.
    • Regularization Parameters: Parameters like alpha in Ridge Regression and C in SVM that control the strength of regularization.

  1. Model Interpretation and Analysis

    • Coefficients in Linear Models: Interpretation of model coefficients in linear regression and classification models.
    • Decision Boundaries: Understanding how decision boundaries change with different parameters in SVM.

Important Equations

  1. Ridge Regression Cost Function:
J(θ)=12m∑i=1m(hθ(x(i))−y(i))2+λ∑j=1nθj2J(\theta) = \frac{1}{2m} \sum_{i=1}^{m} (h_\theta(x^{(i)}) - y^{(i)})^2 + \lambda \sum_{j=1}^{n} \theta_j^2

where λ\lambda is the regularization parameter.

  1. Mean Squared Error (MSE):
MSE=1n∑i=1n(ytrue,i−ypred,i)2\text{MSE} = \frac{1}{n} \sum_{i=1}^{n} (y_{\text{true},i} - y_{\text{pred},i})^2
  1. R² Score:
R2=1−SSRSSTR^2 = 1 - \frac{\text{SSR}}{\text{SST}}

where SSR is the sum of squares of residuals and SST is the total sum of squares.

  1. Precision Score:
Precision=True PositivesTrue Positives+False Positives\text{Precision} = \frac{\text{True Positives}}{\text{True Positives} + \text{False Positives}}
  1. Polynomial Features:
PolynomialFeatures(degree=d)\text{PolynomialFeatures}(degree=d)

generates features of degree up to dd.

  1. KNN Imputation:
KNNImputer(n_neighbors=k,weights="uniform")\text{KNNImputer}(n\_neighbors=k, weights="uniform")

imputes missing values based on the kk nearest neighbors.