Scribd
Upload
29 views

Features Engineering

Feature engineering is the process of using domain knowledge to create features that improve machine learning algorithms. It is fundamental to applying machine learning but also difficult and time-consuming, requiring data scientists to spend 80% of their time on data preparation. Feature engineering increases predictive power and facilitates the machine learning process. Common encoding techniques include one-hot encoding, label encoding, and mean encoding.

Uploaded by

Lovely doll
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
29 views

Features Engineering

Feature engineering is the process of using domain knowledge to create features that improve machine learning algorithms. It is fundamental to applying machine learning but also difficult and time-consuming, requiring data scientists to spend 80% of their time on data preparation. Feature engineering increases predictive power and facilitates the machine learning process. Common encoding techniques include one-hot encoding, label encoding, and mean encoding.

Uploaded by

Lovely doll
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 11

Features Engineering

• Feature engineering is the process of using domain knowledge of the


data to create features that make machine learning algorithms work.
• fundamental to the application of machine learning, and is both
difficult and expensive.
• The need for manual feature engineering can be obviated by
automated feature learning.
Why is Feature Engineering important?
• Increases predictive power of machine learning algorithms
• Facilitates the machine learning process
• Considered an art in data science
• Steps in solving a machine learning problem
Data scientists spend 80% of their time on Data Preparation:

Data scientists spend 80% of their time on Data Preparation:


Types of Encoding Techniques
1. Nominal Encoding : - Where Order of data does not matter
• One Hot Encoding
• One Hot Encoding With Many Categories
• Mean Encoding
2. Ordinal Encoding: — Where Order of data matters
• Label Encoding
• Target Guided Ordinal Encoding
One Hot Encoding:
Label Encoding: Ordinal categories

BE 1

Masters 3

PhD 4

Stats 2
One Hot Encoding With Many Categories

• F1
• 50 categories repeating in most of the features
• Top 10 categories that repeating more (9 coloumns)
Target Guided Ordinal Encoding

Label Feature 1 Output Mean

3 A 1 .73

2 B 1 .4

1 C 0 0.2

…. A 0 …

…. B 0 …
Mean Encoding

You might also like