梯度下降和邏輯迴歸例子(Python程式碼實現)
阿新 • • 發佈:2019-01-14
import numpy as np import pandas as pd import os data = pd.read_csv("iris.csv") # 這裡的iris資料已做過處理 m, n = data.shape dataMatIn = np.ones((m, n)) dataMatIn[:, :-1] = data.ix[:, :-1] classLabels = data.ix[:, -1] # sigmoid函式和初始化資料 def sigmoid(z): return 1 / (1 + np.exp(-z)) # 隨機梯度下降 def Stocgrad_descent(dataMatIn, classLabels): dataMatrix = np.mat(dataMatIn) # 訓練集 labelMat = np.mat(classLabels).transpose() # y值 m, n = np.shape(dataMatrix) # m:dataMatrix的行數,n:dataMatrix的列數 weights = np.ones((n, 1)) # 初始化迴歸係數(n, 1) alpha = 0.001 # 步長 maxCycle = 500 # 最大迴圈次數 epsilon = 0.001 error = np.zeros((n,1)) for i in range(maxCycle): for j in range(m): h = sigmoid(dataMatrix * weights) # sigmoid 函式 weights = weights + alpha * dataMatrix.transpose() * (labelMat - h) # 梯度 if np.linalg.norm(weights - error) < epsilon: break else: error = weights return weights # 邏輯迴歸 def pred_result(dataMatIn): dataMatrix = np.mat(dataMatIn) r = Stocgrad_descent(dataMatIn, classLabels) p = sigmoid(dataMatrix * r) # 根據模型預測的概率 # 預測結果二值化 pred = [] for i in range(len(data)): if p[i] > 0.5: pred.append(1) else: pred.append(0) data["pred"] = pred os.remove("data_and_pred.csv") # 刪除List_lost_customers資料集 # 第一次執行此程式碼時此步驟不要 data.to_csv("data_and_pred.csv", index=False, encoding="utf_8_sig") # 資料集儲存 pred_result(dataMatIn)
如果有錯誤的,請告訴我哈,比較菜!