Which changes in model training would MOST likely improve the model’s F1 score?
(Choose two.)
Continue to use the SageMaker linear learner algorithm. Reduce the number of features with the SageMaker principal component analysis (PCA) algorithm.
Continue to use the SageMaker linear learner algorithm. Reduce the number of features with the scikit-learn multi-dimensional scaling (MDS) algorithm.
Continue to use the SageMaker linear learner algorithm. Set the predictor type to regressor.
Use the SageMaker k-means algorithm with k of less than 1,000 to train the model.
Use the SageMaker k-nearest neighbors (k-NN) algorithm. Set a dimension reduction target of less than 1,000 to train the model.
Explanations:
Reducing the number of features through dimensionality reduction (PCA) may help by simplifying the model, potentially improving generalization and boosting the F1 score.
Multi-dimensional scaling (MDS) is generally used for data visualization rather than feature reduction for improving classification performance, making it unsuitable for this task.
Setting the predictor type to regressor is inappropriate because the objective is classification, not regression. This would not help improve the F1 score.
The k-means algorithm is a clustering technique, not designed for binary classification, which would not directly improve F1 score in this classification scenario.
Using the k-nearest neighbors (k-NN) algorithm with dimensionality reduction can be effective for high-dimensional data, potentially improving classification performance and F1 score.
As I understand it, the answer is:
Continue to use the SageMaker linear learner algorithm. Reduce the number of features with the SageMaker principal component analysis (PCA) algorithm.
I believe the answer is:
Continue to use the SageMaker linear learner algorithm. Reduce the number of features with the SageMaker principal component analysis (PCA) algorithm.