寻找“热点”的算法在 GPS 坐标数据库中
我有大量数据即将放入数据库,它是 GPS 点列表。
我想迭代这个数据库并创建一个“热点”表,其中在一定大小的区域(方形区域或圆形区域 - 我不需要精确)中有大量数据库点)。
谁能推荐可以帮助我解决这个问题的现有算法?
提前致谢!
r3mo
I've got an amount of data that I'm about to put into a database, it's a list of GPS points.
I want to iterate over this database and create a table of 'hot spots' where there are a high number of database points in a certain size of area (either a square area, or a circular area - I don't need to be exact).
Can anyone recommend existing algorithms that might help me with this?
Thanks in advance!
r3mo
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(2)
K 均值聚类将是识别热点的一个很好的起点。请参阅维基百科条目。
K-Means clustering would be a good starting point, for identifying hot-spots. See wikipedia entry.
如何创建具有给定像元大小的栅格并将栅格值分配给每个像素内的点数(密度图)?这是一种基本方法,有一些限制(网格的放置位置和像素大小将影响结果),但如果这就是您所需要的......这可以在 R 中使用 spatstat 包轻松完成。请查看 spatstat 上的此 pdf 教程以获取示例。
除非另一个变量附加到你的点上,否则它不是真正的热点检测,只是点密度的确定......
How about creating a raster with a given cell size and assigning the raster value to the number of points falling within each pixel (a density plot)? It's a basic approach with some limitations (where you place the grid and the pixel size will affect the outcome), but if that's all you need... This could be accomplished easily in R using the spatstat package. Check out this pdf tutorial on spatstat for examples.
Unless another variable is attached to your points, it's not really hotspot detection, just a determination of point density...