DMC Troubleshooting & Problem Resolution 3

Question 1

An ETL job fails because a date column contains values like '31/02/2024'. What is the best resolution?

Accepted Answer

Validate and clean or reject invalid dates during data preparation

Answer

Invalid dates should be detected by validation rules and corrected, flagged, or excluded.

Question 2

After joining two tables, the row count is much larger than expected. What is the most likely cause?

Accepted Answer

Duplicate keys causing a many-to-many join

Answer

Non-unique join keys on both sides multiply rows.

Question 3

A model fails to train because of NaN values in numeric features. Which is an appropriate first step?

Accepted Answer

Impute or remove missing values after analyzing why they are missing

Answer

Missing values must be handled with a method suited to the missingness pattern before training.

Question 4

Distance-based algorithms such as k-NN give poor results when one feature is measured in dollars and another in fractions. What should be done?

Accepted Answer

Scale or normalize the features

Answer

Features on large scales dominate distance calculations unless scaled.

Question 5

Outliers are distorting the mean of a skewed income variable. Which approach is most robust?

Accepted Answer

Use the median or apply a transformation

Answer

The median and log transforms are less sensitive to extreme values.

DMC - Data Mining Certification Practice Test