Next Article in Journal
Expanding Exotic Forest Plantations and Declining Rural Populations in La Araucanía, Chile
Previous Article in Journal
Research on the Distribution and Scale Evolution of Suzhou Gardens under the Urbanization Process from the Tang to the Qing Dynasty
Previous Article in Special Issue
Generative Learning for Postprocessing Semantic Segmentation Predictions: A Lightweight Conditional Generative Adversarial Network Based on Pix2pix to Improve the Extraction of Road Surface Areas
Article

Exploring the Sensitivity of Recurrent Neural Network Models for Forecasting Land Cover Change

Spatial Analysis and Modeling Laboratory, Department of Geography, Simon Fraser University, Burnaby, BC V5A1S6, Canada
*
Author to whom correspondence should be addressed.
Academic Editor: Gunasekaran Manogaran
Land 2021, 10(3), 282; https://doi.org/10.3390/land10030282
Received: 9 February 2021 / Revised: 3 March 2021 / Accepted: 8 March 2021 / Published: 10 March 2021
(This article belongs to the Special Issue Deep Learning Algorithms for Land Use Change Detection)
Recurrent Neural Networks (RNNs), including Long Short-Term Memory (LSTM) architectures, have obtained successful outcomes in timeseries analysis tasks. While RNNs demonstrated favourable performance for Land Cover (LC) change analyses, few studies have explored or quantified the geospatial data characteristics required to utilize this method. Likewise, many studies utilize overall measures of accuracy rather than metrics accounting for the slow or sparse changes of LC that are typically observed. Therefore, the main objective of this study is to evaluate the performance of LSTM models for forecasting LC changes by conducting a sensitivity analysis involving hypothetical and real-world datasets. The intent of this assessment is to explore the implications of varying temporal resolutions and LC classes. Additionally, changing these input data characteristics impacts the number of timesteps and LC change rates provided to the respective models. Kappa variants are selected to explore the capacity of LSTM models for forecasting transitions or persistence of LC. Results demonstrate the adverse effects of coarser temporal resolutions and high LC class cardinality on method performance, despite method optimization techniques applied. This study suggests various characteristics of geospatial datasets that should be present before considering LSTM methods for LC change forecasting. View Full-Text
Keywords: sensitivity analysis; recurrent neural networks; long short-term memory; deep learning; land cover change modelling sensitivity analysis; recurrent neural networks; long short-term memory; deep learning; land cover change modelling
Show Figures

Figure 1

MDPI and ACS Style

van Duynhoven, A.; Dragićević, S. Exploring the Sensitivity of Recurrent Neural Network Models for Forecasting Land Cover Change. Land 2021, 10, 282. https://doi.org/10.3390/land10030282

AMA Style

van Duynhoven A, Dragićević S. Exploring the Sensitivity of Recurrent Neural Network Models for Forecasting Land Cover Change. Land. 2021; 10(3):282. https://doi.org/10.3390/land10030282

Chicago/Turabian Style

van Duynhoven, Alysha, and Suzana Dragićević. 2021. "Exploring the Sensitivity of Recurrent Neural Network Models for Forecasting Land Cover Change" Land 10, no. 3: 282. https://doi.org/10.3390/land10030282

Find Other Styles
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Back to TopTop