Exam AI-900 topic 1 question 62 discussion

I think it should be feature engineering.

Based on this statement from the referenced article: Feature selection: The process of selecting the key subset of features to reduce the dimensionality of the training problem. Feature selection is the answer since feature engineering is the creation of NEW features from raw data to capture additional information not easily apparent in the original feature set.

The answer is Feature engineering! Feature engineering involves transforming raw data into features suitable for machine learning algorithms.   One crucial step is to ensure that numeric variables have a similar scale. This process, often called normalization or standardization, prevents features with larger scales from dominating the model's learning process.   e.g: Imagine you have a dataset with two features: age (ranging from 18 to 65) and income (ranging from $10,000 to $200,000). Without scaling, the income feature might dominate the model's decision-making process, leading to biased results. By scaling both features to a similar range (e.g., 0 to 1), the model can treat both features equally.

Feature Engineering = Creation of NEW features from the given raw data. E.g creation of a "seconds" column from a "days" column. Here we are both normalizing and creating a new feature called "seconds." Feature Engineering is the best answer here. Meanwhile Feature selection is choosing features that have (direct & measurable) impacts on the predictions. E.g. Choosing "Driver's Age" & "Number of Years of Driving" & ignoring "Drivers Names" for predicting the likely hood of having a road accident

Feature Engineering = Creation of NEW features from the given raw data. E.g creation of a "seconds" column from a "days" column. Here we are both normalizing and creating a new feature called "seconds." Feature Engineering is the best answer here.

feature selection is a process of reducing / dropping features, so it is not the correct answer.

This is normalization. In the context of machine learning and data preprocessing, normalization is considered a feature engineering technique, not feature selection.

Normally feature engineering is applied first to generate additional features, and then feature selection is done to eliminate irrelevant, redundant, or highly correlated features.

the given answer is wrong. The correct one is feature engineering

Feature engineering, as it is hinted here: https://learn.microsoft.com/en-us/azure/machine-learning/concept-automated-ml?view=azureml-api-2#feature-engineering

Given answer is wrong. Feature engineering involves transforming or manipulating the input features (variables) in a way that improves the performance or interpretability of a machine learning model. One common technique in feature engineering is scaling or normalizing numeric variables to ensure they have a similar scale or range. When the numeric variables have different scales, it can negatively impact the performance of certain machine learning algorithms. For example, algorithms that are sensitive to the scale of variables, such as gradient descent-based optimization algorithms, may converge slowly or exhibit biased results if the variables have significantly different scales. https://towardsdatascience.com/feature-engineering-for-numerical-data-e20167ec18

Ensuring that the numeric variables in training data are on a similar scale is an example of feature engineering. Feature engineering involves transforming raw data into a format that is more suitable for modeling. It can involve a range of activities, such as dealing with missing values, creating new features from existing ones, encoding categorical variables, and normalizing numeric variables. Normalizing numeric variables, as you mentioned, is a common form of feature engineering to ensure all variables are on a similar scale. This can be particularly important for certain types of models, like neural networks and support vector machines, which can behave poorly if the features are not on similar scales.

Normalization is a **feature engineering** technique used to adjust the values of features in a dataset to a common scale. This is done to facilitate data analysis and modeling, and to reduce the impact of different scales on the accuracy of machine learning models ¹. Is there anything else you would like to know? Source: Conversation with Bing, 4/24/2023 (1) Feature Engineering: Scaling, Normalization and Standardization. https://www.analyticsvidhya.com/blog/2020/04/feature-scaling-machine-learning-normalization-standardization/ Accessed 4/24/2023.

Feature Engineering as per Chat GPT (dealing with data on a similar scale)

feature engineering 100%

Feature Selection is the right answer: Normally feature engineering is applied first to generate additional features, and then feature selection is done to eliminate irrelevant, redundant, or highly correlated features. (https://learn.microsoft.com/en-us/azure/architecture/data-science-process/create-features)

A confusing question (with the given answers) for sure! I am not sure about whether this should be Feature Engineering or Feature Selection