Unit 10: Weka - Practice Quiz

1 What is Weka mainly used for?

2 Weka was developed at which university?

3 Which programming language is primarily used to develop Weka?

4 Which Weka interface provides tabs such as Preprocess, Classify, and Cluster?

5 Which tab in Weka Explorer is used to inspect and prepare a dataset?

6 Which file format is native to Weka?

7 Which common table-based file type can Weka import?

8 Which button in the Preprocess tab is commonly used to load a dataset from the computer?

9 What does an attribute represent in a Weka dataset?

10 What does an instance usually represent in a Weka dataset?

11 Which Weka Explorer tab is used to choose a classification algorithm?

12 Which Weka algorithm builds a decision tree?

13 Which algorithm in Weka is commonly used for clustering?

14 Which factor should be considered when choosing between classification and clustering?

15 Which algorithm is a probabilistic classifier available in Weka?

16 What type of learning is classification?

17 What type of learning is clustering?

18 What does a classification algorithm predict?

19 What is the main purpose of cross-validation in Weka?

20 What does a confusion matrix summarize?

21 A student wants to load a dataset, apply preprocessing filters, train a classifier, and inspect predictions through a graphical interface. Which Weka application is most appropriate?

22 A researcher must compare several algorithms across multiple datasets using repeated runs and statistical significance tests. Which Weka interface best supports this task?

23 In Weka Explorer, which tab should be used to inspect a confusion matrix and evaluate a trained decision tree?

24 A workflow needs to connect data loading, filtering, model training, and evaluation components visually, then save the complete workflow for reuse. Which Weka application should be selected?

25 After installing an additional algorithm through Weka's Package Manager, it does not appear in an already open Explorer window. What is the most appropriate first action?

26 An ARFF file defines outlook as @attribute outlook {sunny,overcast,rainy}. How will Weka interpret this attribute?

27 A CSV column contains the values 1, 2, and unknown. Weka infers the entire column as nominal, but it should be numeric with missing values. What is the best correction before import?

28 A dataset has a categorical target in its first column, but Weka selects the final column as the class by default. What should the user do before training a classifier?

29 A numeric attribute contains missing values represented by ?. Which Weka preprocessing filter can fill them using statistics estimated from the available data?

30 A dataset includes a unique customer ID that has no predictive meaning. What is the most appropriate preprocessing action before classification?

31 A medical dataset contains numeric predictors with very different scales. The user selects IBk with Euclidean distance. Which preprocessing step is most important?

32 A team needs a classification model that can be presented as readable if-then decisions to nontechnical stakeholders. Which Weka algorithm is the best initial choice?

33 A text-classification dataset has thousands of word-frequency attributes and comparatively few training instances. Which algorithm is a reasonable baseline in Weka?

34 A dataset contains labeled examples, and the goal is to predict one of three known species for each new instance. Which type of Weka algorithm is required?

35 An analyst wants a linear maximum-margin classifier and chooses Weka's SMO. The predictors have different numeric ranges. Which setup is most appropriate?

36 A J48 model scores 99% on its training data but only 72% under 10-fold cross-validation. What is the most likely interpretation?

37 For a binary classifier, Weka reports , , and for the positive class. What is the precision for that class?

38 A SimpleKMeans run produces very different clusters when the experiment is repeated with a different random seed. What is the best interpretation?

39 A classifier is evaluated on a highly imbalanced fraud dataset and achieves 98% accuracy by predicting nearly every transaction as legitimate. Which metric should receive greater attention for detecting fraud cases?

40 In Weka Explorer, an analyst has true class labels but wants SimpleKMeans to form clusters without using those labels, while still comparing clusters with the known classes afterward. Which Cluster mode should be used?

41 In Weka Explorer, 10-fold cross-validation produces a J48 tree followed by cross-validation metrics. Which interpretation of this output is correct?

42 A researcher must compare four classifiers over 25 datasets with 10 repeated runs per dataset and perform paired significance tests. Which Weka interface is most appropriate?

43 Consider a command using FilteredClassifier with J48 as its base classifier. What is the purpose of -- before options such as -C 0.1 -M 5?

44 An ARFF relation declares @attribute status {inactive,active}. A sparse instance omits status entirely. How does Weka interpret that attribute?

45 Training data declares @attribute color {red,green,blue}, while a supplied test file declares @attribute color {blue,green,red}. Both files use only these three labels. What should be done?

46 An ARFF attribute is declared as @attribute code {'?','A'}. How should a row distinguish a missing value from the literal nominal value ??

47 A CSV file contains postal codes such as 00123, but automatic loading converts the column to numeric values such as 123. Which intervention best preserves the original information?

48 A text column is imported as Weka's string type, and J48 reports that it cannot handle string attributes. Which pipeline is most appropriate when the text is predictive?

49 A target containing integer values from 0 to 100 is declared numeric in ARFF. A user selects J48 expecting 101 classes. Why is this inappropriate?

50 Twenty classifier and hyperparameter combinations are compared using 10-fold cross-validation, and the best cross-validation score is reported as the final performance estimate. What is the strongest correction?

51 A test set has 100 positive and 9,900 negative cases. Model A predicts every case as negative. Model B produces 20 false negatives and 300 false positives. A false negative costs 100 units and a false positive costs 1 unit. Which choice minimizes stated cost?

52 Records are ordered monthly, and the goal is to predict future months under gradual concept drift. Why can ordinary randomized cross-validation in Weka be misleading?

53 A TF-IDF StringToWordVector filter is fitted once on the complete dataset and then 10-fold cross-validation is run on the resulting attributes. What is the main methodological problem?

54 For a J48 split, an instance has a missing value for the splitting attribute. How does the C4.5-style procedure generally handle that instance?

55 A pruned J48 tree appears to overfit. Holding other options fixed, which change generally increases pruning?

56 Weka displays the following confusion matrix, where rows are actual classes and columns are predicted classes:

40 10 | a

5 45 | b

What are precision and recall for class b?

57 SimpleKMeans uses Weka's default normalized Euclidean distance. If one numeric attribute is converted from dollars to cents by multiplying every nonmissing value by 100, what should occur, ignoring floating-point effects?

58 Two SimpleKMeans runs on the same relation use identical settings except for the random seed and produce different within-cluster sums of squared errors. What is the best explanation?

59 In Weka's Classes to clusters evaluation, a clusterer splits one true class into several pure clusters. Why can the reported error still be substantial?

60 In Weka's EM clusterer, the number of clusters is set to -1. What behavior should be expected?