数据挖掘-Vertibi算法实例及代码

Joker_DDDu

于 2023-04-25 11:18:21 发布

阅读量158

点赞数

分类专栏： ML 文章标签：数据挖掘算法 python

本文链接：https://blog.csdn.net/m0_56863547/article/details/130360767

版权

ML 专栏收录该内容

5 篇文章 0 订阅

订阅专栏

手推示例
Python编程

在这里插入图片描述

code：

import numpy as np


class Vertibi():
    def __init__(self, A, B, pi):
        self.A = A
        self.B = B
        self.pi = pi

    def forward(self):
        self.day1 = np.multiply(self.B[0], self.pi)
        self.path1 = np.array([0, 0])
        self.day2 = np.multiply(np.max(np.multiply(self.day1, self.A), axis=1), self.B[1])
        self.path2 = np.argmax(np.multiply(self.day1, self.A), axis=1)
        self.day3 = np.multiply(np.max(np.multiply(self.day2, self.A), axis=1), self.B[2])
        self.path3 = np.argmax(np.multiply(self.day2, self.A), axis=1)
        self.path = np.array([list(self.path1), list(self.path2), list(self.path3)]).T
        self.weight = np.array([list(self.day1), list(self.day2), list(self.day3)]).T

        return self.weight, self.path

    def back_trace(self):
        # p_star = np.max(self.day3)
        i3_star = np.argmax(self.day3)
        i2_star = self.path[i3_star][2]
        i1_star = self.path[i2_star][1]
        return [i1_star, i2_star, i3_star]


if __name__ == '__main__':
    A = [[0.7, 0.4], [0.3, 0.6]]
    B = [[0.1, 0.6], [0.4, 0.3], [0.5, 0.1]]
    pi = [0.6, 0.4]
    state = ['雨天','晴天']
    mcmc = Vertibi(A, B, pi)
    weight, path = mcmc.forward()
    print("---------------------------矩阵--------------------------")
    print(weight)
    print("\t\t1\t2\t3\t")
    print(f"Rainy:{' '.join(str(round(i,2)) for i in weight[0])}")
    print(f"Sunny:{' '.join(str(round(i,2)) for i in weight[1])}")
    print("----------------------------全部路径--------------------------")
    print(path)
    print("--------------------------回溯最优路径---------------------------")
    print(mcmc.back_trace())
    print(f"预测结果是：{' '.join(state[i] for i in mcmc.back_trace())}")