K-means算法实战-一维数据的聚类

用户有个开关，可以记录用户每次开关灯的状态，并且能够记录开关灯前后的光敏值。一般用户每天开灯时间集中在晚上，关灯时间也基本集中在晚上，但是有一些异常值，比如起夜上厕所开灯、上完厕所再关灯。获取用户开关灯的光敏值若干组，用K-means算法聚类出用户开关灯的阈值。

以开灯的数据为例：

序号开灯前15002550353144605580610073085609401090011400使用K-means算法将数据分为3类，哪个标签的样本数量最多即为阈值：

程序如下：

#include <iostream>#include <vector>#include <cmath>using namespace std;class mL{public:void kmeans(vector<int> &sample,int k){int len = sample.size();vector<int> meanValue(k,0);//初始化均值for(int i = 0;i < k; ++i)meanValue[i] = sample[i];while(1){vector<vector<int> > C(k,vector<int>(k,0));  //用于存储类别vector<int> numC(k,0);//计算每个样本与各个均值的距离for(int i = 0; i < len;++i){int minDis = abs(sample[i] - meanValue[0]);int minDisIndex = 0;for(int j = 1; j < k; ++j){int dis = abs(sample[i] - meanValue[j]);if(dis < minDis){minDis = dis;minDisIndex = j;}}//每个样本属于哪个类C[minDisIndex][numC[minDisIndex]] =  sample[i];numC[minDisIndex]++;}//均值更新int ifBreak = 0;for(int i = 0; i < k;++i){int Sum = 0;for(int j = 0; j < numC[i];++j){Sum += C[i][j];}int lastMeanValue = meanValue[i];meanValue[i] = Sum/numC[i];if(lastMeanValue == meanValue[i])  ifBreak++;}//判断均值是否被更新if(ifBreak == k)  {int maxNum = 0;int maxNumIndex = 0;for(int i = 0;i < k; ++i){if( numC[i] > maxNum){maxNum = numC[i];maxNumIndex = i;}}cout << meanValue[maxNumIndex] << endl;//cout << " Break" << endl;break;}}}};int main(){int arr[] = {500,550,531,460,580,100,30,560,40,900,400};vector<int> lightValue(arr,arr + 10);mL *p = new mL;p->kmeans(lightValue,3);delete p;}