一个神经网络接收外界输入模式时，将会分为不同的对应区域，各区域对输入模式有不同的响应特征，而这个过程是自动完成的。其特点与人脑的自组织特性类似。SOM的目标是用低维（通常是二维或三维）目标空间的点来表示高维空间中的所有点，尽可能地保持点间的距离和邻近关系（拓扑关系）。(When a neural network receives external input mode, it will be divided into different corresponding regions. Each region has different response characteristics to input mode, and this process is completed automatically. Its characteristics are similar to the self - organization characteristics of the human brain. The goal of SOM is to express all points in high dimensional space with the point of low dimensional (usually two-dimensional or three-dimensional) target space, and keep the distance between points and adjacent relations (Topological Relations) as much as possible.)