U2.2 Descriptive Data Mining and Text Mining Quiz

What is the primary goal of dimension reduction in data mining?

• To increase the number of variables
• To replace original variables with new 'meta-variables'
• To remove all variables from the dataset
• To randomly select a subset of variables

Which of the following is NOT a characteristic of principal component analysis?

• It reduces the dimensionality of the data
• It creates weighted combinations of original variables
• It always results in a loss of information
• It maximizes the variance explained by each component

What does the 'elbow method' in k-means clustering help determine?

• The optimal number of clusters
• The best distance measure to use
• The most important variable in the dataset
• The outliers in the data

Which clustering method is more sensitive to outliers?

• k-means clustering
• Hierarchical clustering
• Both are equally sensitive
• Neither is sensitive to outliers

What does the 'lift ratio' in association rules represent?

• The frequency of the antecedent
• The frequency of the consequent
• The effectiveness of the rule compared to random selection
• The total number of transactions

In the context of text mining, what does 'tokenization' refer to?

• Removing punctuation from text
• Converting text to lowercase
• Dividing text into separate terms
• Identifying synonyms in text

What is the purpose of using TFIDF (Term Frequency-Inverse Document Frequency) in text mining?

• To count the total number of words in a document
• To identify the most common words across all documents
• To weight terms based on their frequency and uniqueness
• To remove stopwords from the text

Which distance measure is commonly used in text mining and is unaffected by the magnitude of term frequency?

• Euclidean distance
• Manhattan distance
• Jaccard distance
• Cosine distance

What is a key difference between k-means clustering and hierarchical clustering?

• k-means requires specifying the number of clusters in advance
• Hierarchical clustering cannot handle large datasets
• k-means always produces better results
• Hierarchical clustering doesn't use distance measures

Which of the following is NOT a common method for assessing the quality of clusters?

• Cluster cohesion
• Cluster separation
• Cluster interpretability
• Cluster randomization