Google Professional Machine Learning Engineer
Comments
Loading comments...
Ultimate access to all questions.
As a junior data scientist at a financial services company, you're tasked with developing a logistic regression model to categorize incoming customer text messages into two categories: 'important/urgent' and 'important/not urgent'. The model's performance is critical for timely customer service responses. Given the binary nature of the classification task and the need for probabilistic outcomes between 0 and 1, which loss function is most suitable for evaluating your model's performance? Choose the best option.
A
Mean Square Error (MSE), which is commonly used for regression tasks but may not emphasize the magnitude of errors in probabilistic classifications.
B
Log Loss (Cross-Entropy Loss), which penalizes incorrect classifications more heavily, especially in probabilistic models, making it sensitive to inaccuracies.
C
Mean Absolute Error (MAE), which averages the absolute differences between predicted and actual values but does not significantly emphasize the magnitude of errors.
D
Mean Bias Error (MBE), which calculates the average bias in the predictions but is less effective for classification tasks.
E
Softmax, which is typically used for multi-class classification tasks and not suitable for binary classification.