Auto-ML Tools and Process on Google, Azure and IBM Cloud Platforms
Faculty Mentor
Dr. Hayden Wimmer
Location
Poster 213
Session Format
Poster Presentation
Academic Unit
Department of Information Technology
Background
- The future of Automated Machine Learning (Auto ML) has improved the creativity for data scientists, ML engineers and ML researchers by reducing repetitive tasks in machine learning pipelines.
- Auto ML was designed to effectively solve problems of classification, Multiclass classification, complex systems behavior prediction and selecting unknown parameters that relate the characteristics of complex objects
- The objective of this research study was to predict the different types of network attacks in a network using the different cloud platforms; Google, Azure and IBM cloud platforms.
- We applied the multiclass classification machine learning to train and make predictions of the types of attacks in the UNSWNB15 network security dataset.
- We utilized the following multiclass classifiers, Decision Trees, Random Forest, Gradient Boosting. The evaluating metric was the F1- Score and accuracy.
- The target column or predictor column was in categorical data, then we ran the Term-Transform which converted the categorical data to a Key Numeric type.
Keywords
Allen E. Paulson College of Engineering and Computing Student Research Symposium, Automated Machine Learning, Auto ML, Australian Centre for Cyber Security, ACCS
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Presentation Type and Release Option
Presentation (File Not Available for Download)
Start Date
2022 12:00 AM
January 2022
Auto-ML Tools and Process on Google, Azure and IBM Cloud Platforms
Poster 213
-
The future of Automated Machine Learning (Auto ML) has improved the creativity for data scientists, ML engineers and ML researchers by reducing repetitive tasks in machine learning pipelines.
-
Auto ML was designed to effectively solve problems of classification, Multiclass classification, complex systems behavior prediction and selecting unknown parameters that relate the characteristics of complex objects
-
The objective of this research study was to predict the different types of network attacks in a network using the different cloud platforms; Google, Azure and IBM cloud platforms.
-
We applied the multiclass classification machine learning to train and make predictions of the types of attacks in the UNSWNB15 network security dataset.
-
We utilized the following multiclass classifiers, Decision Trees, Random Forest, Gradient Boosting. The evaluating metric was the F1- Score and accuracy.
-
The target column or predictor column was in categorical data, then we ran the Term-Transform which converted the categorical data to a Key Numeric type.