Google Professional Data Engineer
Get started today
Ultimate access to all questions.
Ultimate access to all questions.
You are working on a regression problem in the domain of natural language processing, utilizing a dataset consisting of 100 million labeled examples. After randomly shuffling the data, you have partitioned it into training and test sets, maintaining a 90/10 ratio respectively. Upon training your neural network model and subsequently evaluating its performance on the test set, you observe that the root-mean-squared error (RMSE) is twice as high on the train set compared to the test set. What approaches can you take to enhance the performance of your model?
Explanation:
The root-mean-squared error (RMSE) of the model being twice as high on the train set as on the test set indicates that the model is underfitting. Underfitting occurs when the model is too simple to capture the underlying pattern in the data, leading to high training error. To address underfitting, it's essential to increase the complexity of the model. This can be done by introducing an additional layer or increasing the size of vocabularies or n-grams used. Therefore, the correct answer is D.