Answers are grounded in this paper and cite their sources.
With the continuous improvement of sensor technology, machine learning and data processing capabilities, underwater navigation gradually turns to the data-driven model. In this paper, we make full use of the data in the appendix to make a reasonable judgment on the classification and prediction of underwater adaptive areas, and evaluate the performance. K-means clustering method is used to cluster the characteristics of gravity anomalies in different regions. When the number of clustering centers is set from 2 to 11:00, the clustering results show that when the number of clusters is 2, 3, 5, 7, 8, 9, all 100 iterations converge, and the clustering effect is good. When the number of clustering is 2, 3 and 5, the sum of the best distance is larger, and the distribution of contour coefficient has negative expansion; when the number of clustering is 7 and 8, the distribution of contour coefficient has negative expansion; when the number of clustering is 9, the contour coefficient is larger, the distribution is uniform on the front, the part of the contour coefficient less than 0 is very small, the convergence in the iterative process, the optimal distance and the sum of the optimal distance is relatively small: 63711.2. Therefore, when the number of clusters is 9, the fitting effect of each region is the best. Mark subclass 1 as area 1, mark subclass 2 as area 2, and so on.
No takes yet. Share an insight, caveat, or question.
Xiangyu Sun (2023) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: