Question

Which of the following statements describes an overfitted model the best. a. Performs worse on training data as the training process proceeds, while performing better on a held-out test data b. Its performance doesn't improve by additional training neither on the training data nor on the test data c. Performs better on training data as the training process proceeds, while performing worse on a held-out test data d. Performs worse on both the training data and the test data as the training process proceeds.

Answer 1

c. Performs better on training data as the training process proceeds, while performing worse on a held-out test data

Step-by-step explanation:

An over-fitted model is one that will perform best on training but would fail or do worse on a held-out test data.

Such models are optimum for a just a particular set of data but would grossly failed when extrapolated to some other data set not novel to it.

• Over-fitting a model implies that a model closely corresponds to a set of data but would not perform well with others.
• It is usually as a result of a model adapting the noise and other details of a particular data set and thereby incorporates it.
• This makes it difficult for the model to fit into another data set.