如何在python中可视化/绘制SOM？

Question

我想为我的群集结果(this kind of map)创建一个地图。例如，这是我使用SOM生成的结果。

我输入的例子（基于DNA序列基序）：

主题1 = 0.19,0.95,0.01,0,0.76,0,1.04,0,0,0.05,0,1,0,0,1,0

主题2 = 0,0,0,0,0,0,1,0.3,0.05,0.15,0.7,0.6,0.05,1.15,0.2,0.8

主题3 = 0.9,0,0,1.1,0,0,0,0.45,0.035,0,0.015,0.15,1.665,0,0.335,1.35

主题4 = 1,0,0,1.16,0.036,0,0.0032,0.4,0.294,0,0.025,0.04,1.5888,0.04,0.371,1.04

输出（在python中使用SOM运行）： 培训输入集群： motif 1 = Cluster 1

主题2 =群集2

motif 3 = Cluster 1

主题4 =群集1

节点1的权重： 1.366,0.951,0.819,0.919,0.812,0.688,0.802,0.622,0.999,0.574,0.618,0.803,0.880,0.721,0.741,0.963 节点2的权重： 1.366,0.951,0.819,0.919,0.812,0.688,0.802,0.622,0.999,0.574,0.618,0.803,0.880,0.721,0.741,0.963

Answer 1

我一直在寻找相同的产品，但是遇到了一个不错的选择。它被称为“星爆图”，并在Lutz Hamel博士及其前任学生开发和维护的名为popsom的SOM实现中。

Example of the graphical representation:

popsom for python

popsom for R