Data for machine learning · Question 38 of 60
Why should the operational data match the characteristics of the training data?
1 answer is correct. Medium
Show answer
Answer
- D So that the model performs as expected in production, for example regarding distributions and ranges
Explanation
For the model to work as well in operation as in testing, the real data should resemble the training data, for example in its distribution and the value ranges of the features. If it differs greatly, the assumptions about performance and safety may no longer be true. The individual preparation steps, on the other hand, may be rebuilt so that they run faster or scale better.