AIF-C01 Question 42
Single answerA company is building an ML pipeline to predict customer churn. The team has collected the data and conducted exploratory data analysis (EDA). What should the team do next to ensure the data is in a suitable format for training the model?
- A
Perform data pre-processing to handle missing values, normalize features, and address outliers
- B
Deploy the model to a production-ready endpoint for inference
- C
Conduct hyperparameter tuning to optimize the model's performance
- D
Evaluate the model's performance on a validation dataset
Show answer and explanation
Correct answer: A
Explanation
After conducting exploratory data analysis, the next step in the ML pipeline is to perform data pre-processing. This step ensures that the data is clean and consistent by addressing issues like missing values, feature scaling, and outliers. It is a critical step to prepare the data for model training, which occurs later in the pipeline.
- A. Correct.
Correct: Data pre-processing is the next logical step after EDA. It ensures the data is clean, consistent, and ready for the model training phase.
- B. Incorrect.
Incorrect: Deployment happens after the model has been trained, evaluated, and finalized. The team is not yet at this stage.
- C. Incorrect.
Incorrect: Hyperparameter tuning is part of the model training phase, which typically comes after data pre-processing and feature engineering.
- D. Incorrect.
Incorrect: Model evaluation is performed after training the model, not before pre-processing the data.