Decision Tree Ensemble vs. N.N. Deep Learning : efficiency comparison for a small image dataset

Treboux, Jérôme (University of Applied Sciences and Arts Western Switzerland (HES-SO Valais-Wallis)) ; Genoud, Dominique (University of Applied Sciences and Arts Western Switzerland (HES-SO Valais-Wallis)) ; Ingold, Rolf (University of Fribourg, Switzerland)

This paper presents a study of the efficiency of machine learning algorithms applied on an image recognition task. The dataset is composed of aerial GeoTIFF images of 5 different vineyards taken with a drone. It presents the application of two different classification algorithms with an efficiency comparison over a small dataset. A Neural Network algorithm for classification through the TensorFlow platform will be explained first, and a Decision Tree Ensemble algorithm for classification through a machine learning platform will be explained second. This work shows that the accuracy of the Decision Tree Ensemble algorithm (94.27%) outperforms the accuracy of the Deep Learning algorithm (91.22%). This result is based on the final detection accuracy as well as on the computation time.


Conference Type:
full paper
Faculty:
Economie et Services
School:
HEG-VS
Institute:
Institut Informatique de gestion
Subject(s):
Informatique
Publisher:
Jakarta, Indonesia, 12-13 May 2018
Date:
2018-05
Jakarta, Indonesia
12-13 May 2018
Pagination:
6 p.
Published in:
Proceedings of International Workshop on Big Data and Information Security (IWBIS 2018)
Appears in Collection:



 Record created 2018-11-16, last modified 2019-11-28

Fulltext:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)