UTpaper

Free mock tests & test series

Question 94Asked in BEL PE Computer Science, 20 Dec 2025 (5:00 PM)

A data scientist is developing a predictive model for loan default detection using a dataset that contains labelled examples of previous loan applications…

A data scientist is developing a predictive model for loan default detection using a dataset that contains labelled examples of previous loan applications marked as 'defaulted' or 'not defaulted'. However, during model training, the algorithm shows very high accuracy on training data but performs poorly on new unseen test data. Despite adding more labelled data, the issue persists. Which of the following best explains the underlying problem in the supervised learning process?
  1. AThe model suffers from overfitting due to poor generalisation.
  2. BThe dataset lacks sufficient labelling for supervised learning.
  3. CThe model applies an unsupervised clustering technique instead of classification.
  4. DThe algorithm fails due to lack of a defined loss function.
Show answerHide answer

Correct answer: (A) The model suffers from overfitting due to poor generalisation.

From BEL PE Computer Science PYQ, 20 Dec 2025 (5:00 PM)

This is question 94 of BEL PE Computer Science PYQ, 20 Dec 2025 (5:00 PM). See all questions of this paper with answers, or attempt the paper as a timed online test to get your score and all-India rank.