UTILIZING FEATURETOOLS IN AUTOMATICALLY CREATING FEATURE ENGINEERING
Back in the time when the technological knowledge has bloom into the 21st century, technology has become one of the solutions that have been focused especially in using the machine learning to help the human making a better decision making. In the machine learning, there are feature engineerin...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Year Project |
| Language: | English |
| Published: |
IRC
2020
|
| Subjects: | |
| Online Access: | http://utpedia.utp.edu.my/21727/1/24646_Muhamad%20Amir%20Izzat%20Bin%20Azmi.pdf http://utpedia.utp.edu.my/21727/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Back in the time when the technological knowledge has bloom into the 21st century,
technology has become one of the solutions that have been focused especially in
using the machine learning to help the human making a better decision making. In
the machine learning, there are feature engineering process where this method has
evolved extensively in construction of novel features from the data provided within
the goals to improvise the predictive learning performance. This process has been
performed manually because it relies on the human domain knowledge as it a time�consuming factor that are used during the project of data science workflow.
In this project, presence of the framework called Featuretools helps to
automatically perform feature engineering a set of related tables. The open-source
Python library explores the various feature construction choices based on the method
known as Deep Feature Synthesis. Additionally, the deep feature synthesis stacks of
multiple transformation and aggregation operation called Feature Primitives, to
create features from data spread across many tables. In the other hand, the system
allow user to specify domain or data specific choices to prioritize the exploration.
The implementation of automation on feature generation was a success.
Using the concept can perform deep feature synthesis to create new features and
functions applied to one or more columns in a single table or to build new features
from multiple tables. The output for the project is to obtain the recognition of
utilizing automated feature engineering with features compare to the manual way for
the data analysis and machine learning pipelines. |
|---|