335x Filetype PDF File size 1.87 MB Source: www.sto.nato.int
BathEMADE: Evolutionary Multi-
objective Algorithm Design
Engine for Bathymetric LIDAR
Jason Zutty
Rodd Talebi
James Rick
Christopher Valenta
Domenic Carr
1 Gregory Rohling
2
How is a Machine Learning Algorithm Made?
• Involves a number of steps
Data Feature Model Model Parameter Model
Raw Data Preparation Engineering Selection & Evaluation Optimization Deployment
Training
2
3
How is a Machine Learning Algorithm Made?
• Involves a number of steps
Data Feature Model Model Parameter Model
Raw Data Preparation Engineering Selection & Evaluation Optimization Deployment
Training
• Each step represents a choice of one method from a large library of options
• Each choice has a number of parameters that affect the performance
• Overall this is a time consuming, iterative, and labor intensive process
3
4
How is a Machine Learning Algorithm Made?
• Involves a number of steps
Data Feature Model Model Parameter Model
Raw Data Preparation Engineering Selection & Evaluation Optimization Deployment
Training
PCA
FFT
Wavelet
Transform
Scaling
Recursive
Feature
Elimination
4
no reviews yet
Please Login to review.