Scribd
Upload
8 views

DM Lab Assignment 2

Lab

Uploaded by

Memoona Ishfaq
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
8 views

DM Lab Assignment 2

Lab

Uploaded by

Memoona Ishfaq
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

COMSATS UNIVERSITY ISLAMABAD

Department of Computer Science


Lab Assignment 2

Course: Data Mining (DSC306) Total marks 10

[CLO 2 Apply classification and clustering techniques such as decision trees, rule-based classifiers, K-means to real-
world datasets.]

Question 1

Task Description:

• Train at least 3 classifiers on a dataset and compute the accuracy, precision, recall, and F1-score for
each classifier.
• Visualize the ROC curve for each classifier and compare the AUC values.
• Perform 5-fold cross-validation on the classifiers and record the results.

Question 2

Comprehensive k-NN Evaluation

Task Description:

• Dataset: Choose any dataset of your choice (can be from UCI repository or other sources).
• Implement k-NN: Apply k-NN to classify the data using at least two different distance metrics.
• Optimize k: Experiment with different values of k and select the one that yields the best accuracy.
• Visualization: If possible, visualize the decision boundaries for a subset of features.
• Submission: Submit the code, a report analyzing your findings (distance metrics, optimal k, decision
boundaries), and the accuracy results.

Question 3

End-to-End Classifier Implementation and Evaluation


Task Description:
• Dataset: Choose a dataset of your choice (e.g., from the UCI Machine Learning Repository).
• Objective: Implement both a Naive Bayes classifier and a Neural Network on the dataset. Tune the
hyperparameters for both classifiers to achieve the best results.
• Performance Comparison: Compare the two classifiers in terms of accuracy, precision, recall, F1-score,
and training time.
• Submission: Submit your code, results, and a short report analyzing the performance of the two models
and discussing why one might perform better than the other.

Question 4

End-to-End SVM Implementation and Analysis


Task Description:
• Dataset: Choose a dataset from the UCI Machine Learning Repository or Kaggle.
• Objective: Implement SVM on this dataset using both linear and non-linear kernels. Perform
hyperparameter tuning using Grid Search to optimize the performance of your model.
• Analysis: Compare the performance of different kernels (linear, RBF) and hyperparameters. Provide a
report on the best model, its accuracy, precision, recall, F1-score, and decision boundary (if applicable).
• Submission: Submit the code, results, and the analysis report.

---------------------------------------------

You might also like