Share this post on:

. All experiments are carried out on a laptop with Windows ten operating
. All experiments are carried out on a computer with Windows 10 operating method, NVIDIA RTX 2060 Super GPU, and 64 GB RAM. The overall accuracy (OA), typical accuracy (AA), and kappa coefficient (Kappa) are adopted as the evaluation criteria. Different proportions of training, validation, and testing samples for every dataset are utilized to confirm the effectiveness from the proposed model thinking about the unbalanced categories in 4 benchmarks. The batch size and epochs are set to 16 and 200, IQP-0528 HIV respectively. Stochastic gradient descent (SGD) is adopted to optimize the coaching parameters. The initial learning price is 0.05 and decreases by 1 each and every 50 epochs. Each of the experiments are repeated 5 instances to prevent errors. three.3. Evaluation of Parameters (1) Impact of Principal Component: In this section, the influence with the variety of principal elements C is tested on classification final results. PCA is first employed to decrease the dimensionality of your bands to 20, 30, 40, 50, and 60, respectively. The experimental final results on 4 datasets are shown in Figure 5. For the University of Pavia and Kennedy Space Center datasets, the values of OA, AA, and Kappa rise from 20 (PU_OA = 98.81 , KSC_OA = 96.92 ) and attain a peak at 30(PU_OA = 98.96 , KSC_OA = 99.07 ). The enhance in OA values around the KSC dataset is a lot greater than that around the PU dataset. It may be observed that the number of principal elements has a substantial effect on the KSC dataset. When the principal element bands exceed 30, these indicators decline to vary degrees. Even though for the Salinas Valley and GRSS_DFC_2013, the values of OA, AA, and Kappa seem to possess no such relationships together with the principal elements. The OA values fluctuate in various variety of principal components. The phenomenon is probably caused by the fact that the latter two datasets have a higher land-cover resolution but a reduced spectral band sensitivity.Micromachines 2021, 12, x FOR PEER Critique Micromachines 2021, 12,9 of 17 9 ofUniversity of PaviaOA AA KappaKennedy Space CenterOA AA KappaAccuraciesAccuracies20 30 40 5090 20 30 40 50principal componentsprincipal components(a)99.(b)OA AA Kappa 98.Salinas ValleyGRSS_DFC_OA AA Kappa97.8 99.AccuraciesAccuracies20 30 40 5097.99.97.99.2 97.99.97.0 20 30 40 50principal componentsprincipal elements(c)(d)Figure five. OA, AA, and Kappa accuracies with various principal components on 4 datasets. (a) Effect of principal Figure 5. OA, AA, and Kappa accuracies with diverse principal components on 4 datasets. (a) Effect of principal components on University of Pavia dataset, (b) Effect of principal components on Kennedy Space Center, (c) Impact of elements University Impact of principal components (c) Impact of principal elements on Salinas Valley dataset, (d) Effect of principal components on GRSS_DFC_2013 dataset. principal components on Salinas Valley dataset, (d) Impact of principal components on GRSS_DFC_2013 dataset.(two) Influence of Spatial Size: The GLPG-3221 Membrane Transporter/Ion Channel option of your spatial size of your input image block features a (two) Impact of Spatial Size: The option from the spatial size of your input image block has crucial influence on classification accuracy. ToTo uncover the top spatial size,is necessary to a essential influence on classification accuracy. uncover the best spatial size, it it really is necessary test the model byby adopting diverse spatial sizes: C 9 9, C 1 11, C 13 13, to test the model adopting distinctive spatial sizes: C 9 9, C 11 11, C 13 13, C 15 15, C 17 17, and C 19 19, where.

Share this post on:

Author: haoyuan2014