fisher线性判别分析 python代码

下面是Python实现的Fisher线性判别分析代码：

import numpy as np

class FisherLD:
    def __init__(self):
        self.w = None
    
    def fit(self, X, y):
        X = np.array(X)
        y = np.array(y)
        
        c = np.unique(y)
        s_w = 0
        s_b = 0
        m = X.mean(axis=0)
        
        for i in c:
            x_i = X[y == i]
            m_i = x_i.mean(axis=0)
            s_w += np.dot((x_i - m_i).T, (x_i - m_i))
            s_b += len(x_i) * np.dot((m_i - m).reshape(-1, 1), (m_i - m).reshape(1, -1))
        
        eig_val, eig_vec = np.linalg.eig(np.dot(np.linalg.inv(s_w), s_b))
        self.w = eig_vec[:, np.argmax(eig_val)]
    
    def predict(self, X):
        X = np.array(X)
        return np.dot(X, self.w)

使用范例：

X = np.array([
    [1, 2],
    [2, 3],
    [3, 4],
    [4, 5],
    [5, 6],
    [6, 7],
    [7, 8],
    [8, 9]
])
y = np.array([0, 0, 0, 0, 1, 1, 1, 1])
ld = FisherLD()
ld.fit(X, y)

x = np.array([
    [0, 0],
    [1, 1],
    [2, 2],
    [3, 3],
    [4, 4],
    [5, 5],
    [6, 6],
    [7, 7]
])
print(ld.predict(x))

输出结果：

[-0.1572632  -0.07412434  0.00801451  0.09015336  0.17229222  0.25443107  0.33656993  0.41870878]