Microsoft Azure AI Fundamentals (AI-900) Practice Exam

What process involves ensuring that numeric variables in training data are on a similar scale?

1. Feature extraction

2. Feature selection

3. Data normalization

4. Data augmentation

The correct answer is: Feature selection

The process that involves ensuring that numeric variables in training data are on a similar scale is data normalization. Normalization is a crucial technique in preparing data for machine learning, as it helps to standardize the range of independent variables or features of the data. By scaling these numeric variables to a common scale, such as between 0 and 1 or transforming them to have a mean of 0 and a standard deviation of 1, normalization improves the training process. This is particularly important for algorithms that rely on the distance between data points, such as k-nearest neighbors and gradient descent optimization methods, because it prevents features with larger numeric ranges from disproportionately influencing the model. Feature extraction refers to the process of transforming raw data into a set of usable features for modeling, while feature selection focuses on choosing the most relevant variables from the dataset to improve model performance. Data augmentation involves creating new training samples from existing data, typically used in scenarios like image processing to enhance the dataset by varying attributes like rotation or scaling. None of these processes directly address the scaling of numeric variables, making data normalization the key technique in this context.