All MCQs

Each transaction contains a set of purchased items.
This representation supports frequent itemset and association rule mining.

A data matrix is the standard tabular representation used by many algorithms.
Each row is an object and each column is a feature.

For asymmetric attributes, the presence state is often more informative than absence.
Measures such as Jaccard similarity ignore shared zeros for this reason.

Binary attributes take two values such as yes/no or 0/1.
They may be symmetric or asymmetric depending on whether both states are equally important.

Ratio variables have equal intervals and an absolute zero point.
Examples include weight, distance, and transaction amount.

Interval scales support meaningful differences but not meaningful ratios.
Temperature in Celsius is a common example because zero is arbitrary.

Ordinal values have a meaningful ranking such as low, medium, and high.
The distance between adjacent categories is not necessarily equal.

Nominal values distinguish categories without ranking them.
Examples include color, country, and product type.

Semi-supervised methods exploit unlabeled structure while learning from limited labels.
This is useful when labeling is expensive but raw data are abundant.

The historical labels provide target outcomes for model training.
The model learns to predict the known class for future transactions.

The problem has no predefined customer classes.
Clustering can discover groups with similar purchasing or engagement patterns.

Anomaly detection searches for rare or unusual cases.
It is widely used in fraud detection, intrusion detection, and equipment monitoring.