Feature Engineering

Question 1

What is the 3 main goal of the analyze stage in machine learning?

Answer 1

1. Understanding response variables and how they’re structured. continuous? categorical?
2. Explore predictor variables.
3. Featuring Engineering

Question 2

What is the 3 general categories of Feature Engineering?

Answer 2

feature selection, feature extraction, feature transformation.

Question 3

Explain the concept of feature engineering and its importance in machine learning.

Answer 3

Feature engineering involves selecting, transforming, and creating relevant features from raw data to improve model performance. It plays a crucial role in enhancing model accuracy.

Question 4

What is feature engineering in machine learning

Answer 4

Feature engineering is the process of using practical, statistical, and data science knowledge to select, transform, or extract characteristics, properties, and attributes from raw data for the construction of machine learning models.

Question 5

How does feature engineering improve model’s performance

Answer 5

• by solving the data structure issues.
• A well-structured data can help the model to detect predictive signals better.

Question 6

Is feature engineering dependent on the type of data used

Answer 6

Yes, the process of feature engineering is highly dependent on the type of data you’re working with.

Question 7

What is the difference between feature engineering and Exploratory Data Analysis (EDA)

Answer 7

• Feature engineering goes a step beyond EDA.
• EDA involves exploring the data
• feature engineering involves selecting, extracting, or transforming variables or features from datasets for the construction of machine learning models.

Question 8

What is the goal of feature selection in data engineering

Answer 8

to select the features in the data that contribute the most to predicting your response variable. This usually involves dropping features that do not help in making a prediction.

Question 9

What does feature transformation involve in machine learning

Answer 9

• modifying the existing features
• to improve accuracy when training the model.
• This could involve changing the data from numerical to categorical or creating new categories based on the data.

Question 10

Can you give an example of feature transformation

Answer 10

For instance, if your data includes exact temperatures, but you only need a feature that indicates if it’s hot, cold, or temperate. To make that transformation, you could define some cut off points for the data, such as defining anything above 80°F as hot, anything below 70 as cold, and anything in between as temperate.

Question 11

What is feature extraction in machine learning

Answer 11

taking multiple features to create a new one to improve the accuracy of the algorithm.

For example, creating a new variable that becomes true if the temperature is warm and the humidity is high, and false otherwise.

Question 12

How does feature extraction benefit machine learning models

Answer 12

by creating new features that capture important information in a format that’s more understandable for the model, potentially leading to improved accuracy.

Question 13

What is Feature Selection in machine learning

Answer 13

Feature Selection is the process of picking variables from a dataset that will be used as predictor variables for a model.

The goal is to find the predictive and interactive features and
- exclude redundant and irrelevant features
- to improve model performance.

Question 14

What are the three types of features in Feature Selection

Answer 14

• Predictive
• Interactive
• Irrelevant

Question 15

How does Feature Selection fit into the PACE workflow

Answer 15

Feature Selection occurs at multiple stages of the PACE workflow.

• Plan phase, where you define your problem and decide on a target variable to predict. It occurs again during the
• Analyze phase, where after exploratory data analysis, it might be clear that some features might not be suitable for modeling.
• Construct phase, where the goal is to find the smallest set of predictive features that still results in good overall model performance.

Question 16

Why is Feature Selection important in model simplicity and explainability

Answer 16

• Models with fewer, but more predictive features are simpler - simpler models are generally more stable and easier to understand.
• Hence, data professionals often base final model selection not solely on score, but also on model simplicity and explainability.

Question 17

How do data professionals perform Feature Selection during the Construct phase

Answer 17

They typically use statistical methodologies to determine which features to keep and which to drop. This could be as simple as ranking the model’s feature importance and keeping only the top percentage of them, or keeping the top features that account for a certain percentage of the model’s predictive signal.

Question 18

Predictive Feature

Answer 18

Features that by themselves contain information useful to predict the target

Question 19

Interactive Feature

Answer 19

Features that are not useful by themselves to predict the target variable, but become predictive in conjunction with other features

Question 20

Irrelevant Feature

Answer 20

Features that don’t contain any useful information to predict the target

Question 21

What is a redundant feature in machine learning

Answer 21

A redundant feature is a feature that is highly correlated with other features and therefore does not provide the model with any new information.

Question 22

Can predictive features also be redundant and when?

Answer 22

Yes, predictive features can also be redundant if they are highly correlated with other features.

Question 23

What is the goal of feature selection

Answer 23

The goal of feature selection is to find the predictive and interactive features while excluding redundant and irrelevant features.

Question 24

What is feature transformation?

Answer 24

Feature transformation is a process where you take existing features in the dataset, and alter them so that they’re better suited to be used for training the model.

Question 25

What is log normalization and when is it used

Answer 25

Log normalization is a type of feature transformation used to handle continuous variables with skewed distributions. It involves taking the log of a skewed feature, reducing the skew and making the data better for modeling.

Question 26

What is scaling in the context of feature transformation

Answer 26

Scaling is a type of feature transformation where you adjust the range of a feature’s values by applying a normalization function to them. This helps prevent features with very large values from having undue influence over a model compared to features with smaller values but which may be equally important as predictors.

Question 27

What is the difference between normalization and standardization in feature scaling

Answer 27

• Normalization transforms data to fall within the range [0, 1]
• Standardization, on the other hand, transforms each value within a feature so they collectively have a mean of zero and a standard deviation of one.

Question 28

What is encoding in feature transformation

Answer 28

Encoding is a form of feature transformation that involves the process of converting categorical data to numerical data. This enables machine learning models to interpret and process the categorical data.

Question 29

What is the benefit of scaling in feature transformation?

Answer 29

This helps prevent features with very large values from having undue influence over a model compared to features with smaller values but which may be equally important as predictors.

Question 30

What is feature extraction in machine learning?

Answer 30

• involves producing new features from existing ones rather than simply changing one that already exists.
• with the goal of having features that deliver more predictive power to your model.

Question 31

How can feature extraction improve a machine learning model?

Answer 31

• by providing more relevant and insightful data for the model to learn from.
• By transforming or combining existing features, we might reveal hidden relationships or patterns that help the model make more accurate predictions.

Question 32

Can you provide an example of feature extraction

Answer 32

Yes, consider a feature called “Date of Last Purchase,” which contains information about when a customer last purchased something from the company. Instead of giving the model raw dates, a new feature can be extracted called “Days Since Last Purchase.” This could tell the model how long it has been since a customer has bought something, giving insight into the likelihood that they’ll buy something again in the future.

Question 33

Can features be extracted from multiple variables

Answer 33

Yes, features can also be extracted from multiple variables. For example, there are two variables: “Days Since Last Purchase” and “Price of Last Purchase.” A new variable could be created from these by dividing the price by the number of days since the last purchase, creating a new variable altogether.

Question 34

What is the key difference between feature selection, feature transformation and feature extraction?

Answer 34

Feature Selection = dropping any and all unnecessary or unwanted features from the dataset
Feature Transformation = editing features into a form where they’re better for training the model
Feature Extraction = creating brand new features from other features that already exist in the dataset

Question 35

Why is feature transformation important to consider?

Answer 35

Many types of models are designed in a way that requires the data coming in to be numerical.
So, transforming categorical features into numerical features is an important step.

Question 36

How to use scaler?

Answer 36

you must fit it to the training data and transform both the training dataandthe test data using that same scaler.

Import the scaler function
from sklearn.preprocessing import MinMaxScaler

Instantiate the scaler
scaler = MinMaxScaler()

Fit the scaler to the training data
scaler.fit(X_train)

Scale the training data
X_train = scaler.transform(X_train)

Scale the test data
X_test = scaler.transform(X_test)