
Ultimate access to all questions.
In the context of Exploratory Data Analysis (EDA) for a machine learning project aimed at predicting customer churn for a telecom company, the team has identified that the raw data contains customer usage patterns, service complaints, and demographic information. The project has constraints including limited computational resources and the need for the model to be interpretable by non-technical stakeholders. Given these constraints and objectives, what is the primary goal of feature engineering in this scenario? Choose the best option.
A
To directly deploy the machine learning model into production without further adjustments.
B
To create visually appealing data representations for stakeholder presentations.
C
To remove all missing values and outliers from the dataset to simplify the analysis.
D
To transform and create new features from the existing data that enhance the model's predictive performance while ensuring the features are interpretable.
E
Both to visualize the data in a way that highlights customer churn patterns and to clean the data by removing irrelevant features.