本文主要向大家介绍了数据科学第一战：Iris的Python语言数据分析，通过具体的内容向大家展示，希望对大家学习Python语言有所帮助。

用pyhton具体实现机器学习算法感觉并没有Octave方便

幸好python有专门的scikit-learn库来直接用，所以以后的策略就是用Octave学习和理解算法

用python应用库算法解决Kaggle问题

1，Iris数据集逻辑回归Python版，数据地址:https://www.kaggle.com/chuckyin/iris-datasets

import numpy as npimport pandas as pdfrom sklearn.linear_model import LogisticRegressionfrom sklearn.cross_validation import train_test_split

df = pd.read_csv('iris.csv')

y = df['Species']

X = df.drop('Species', axis=1)

x_train, x_test, y_train, y_test = train_test_split(X, y, test_size = 0.2)

clf = LogisticRegression()

clf.fit(x_train, y_train)

score = clf.score(x_test, y_test)print(score)

 分类器的正确率只有93.3%，这还是没有把数据集分为训练集和测试集，接下来分20%的数据集为测试集，80%的数据集为训练集

 1 import numpy as np 2 import pandas as pd 3 from sklearn.linear_model import LogisticRegression 4  5  6 df_train = pd.read_csv('iris_train.csv') 7 df_test = pd.read_csv('iris_test.csv') 8 y_train = y = df_train['Species'] 9 x_train = X = df_train.drop('Species', axis=1)10 y_test = y = df_test['Species']11 x_test = X = df_test.drop('Species', axis=1)12 clf = LogisticRegression()13 clf.fit(x_train, y_train)14 15 score = clf.score(x_test, y_test)16 print(score)

正确率只有66.7%，数据太少，只有150个样例，4个特征，接下来换神经网络试试

import numpy as np
import pandas as pd

def sigmoid(z):

    return 1 / (1 + np.exp(-z))

def sigmoidGradient(z):

    return (1 - sigmoid(z)) * sigmoid(z)

def predict(X, theta1, theta2):

    # forwardPropagation

    m = len(X)

    a1 = np.hstack((np.ones([m, 1]), X))

    z2 = np.dot(a1, theta1)

    a2 = sigmoid(z2)

    a2 = np.hstack((np.ones([m, 1]), a2))

    z3 = np.dot(a2, theta2)

    # print (z3)

    a3 = sigmoid(z3)

    return a3

def convert(y):

    t = y.values

    sp = np.unique(t)

    tt = np.zeros([len(t), len(sp)])

    for i in range(len(t)):

        for j in range(len(sp)):

            if t[i] == sp[j]:

                tt[i][j] = 1

    return tt

def getData(fileName):

    df_train = pd.read_csv(fileName)

    y = convert(df_train['Species'])

    X = df_train.drop(['Id', 'Species'], axis=1).values

    return X, y

X_train, y_train = getData('iris_train.csv')

X_test, y_test = getData('iris_test.csv')

np.random.seed(1)

theta1 = 2 * np.random.rand(5, 4) - 1

theta2 = 2 * np.random.rand(5, 3) - 1

for iter in range(1, 20000):

    # forwardPropagation

    m = len(X_train)

    a1 = np.hstack((np.ones([m, 1]), X_train))

    z2 = a1.dot(theta1)

    a2 = sigmoid(z2)

    a2 = np.hstack((np.ones([m, 1]), a2))

    z3 = a2.dot(theta2)

    a3 = sigmoid(z3)

    # backPropagation

    delta3 = (a3 - y_train).T

    if (iter % 10000) == 0:

        print("Error:" + str(np.mean(np.abs(delta3))))

    grad2 = np.dot(a2.T, delta3.T)

    delta2 = np.dot(theta2[1:, ], delta3) * sigmoidGradient(z2.T)

    #delta2 = np.dot(theta2[1:, ], delta3) * (1 - a2.T[1:, ]) * a2.T[1:, ]

    grad1 = np.dot(a1.T, delta2.T)

    theta1 -= grad1 / m

    theta2 -= grad2 / m

res = np.argmax(predict(X_test, theta1, theta2), axis=1)

print(np.mean(np.argmax(y_test, axis = 1) == res))

还是神经网络强大，100%的准确率，当然也是因为数据偏少，偏差值收敛，每迭代10000次的偏差值

Error:0.013312372435
Error:0.0119592971635
Error:0.0115683980319

加上正则化，准确率准确率虽然没有降低(也可能是数据少)， 但是偏差值发散，每迭代10000次的偏差值

Error:0.0491758284265
Error:0.0457085051291
Error:0.0523013950821

import numpy as npimport pandas as pd

def sigmoid(z):

    return 1 / (1 + np.exp(-z))

def sigmoidGradient(z):

    return (1 - sigmoid(z)) * sigmoid(z)

def predict(X, theta1, theta2):

    # forwardPropagation

    m = len(X)

    a1 = np.hstack((np.ones([m, 1]), X))

    z2 = np.dot(a1, theta1)

    a2 = sigmoid(z2)

    a2 = np.hstack((np.ones([m, 1]), a2))

    z3 = np.dot(a2, theta2)

    # print (z3)

    a3 = sigmoid(z3)

    return a3

def convert(y):

    t = y.values

    sp = np.unique(t)

    tt = np.zeros([len(t), len(sp)])

    for i in range(len(t)):

        for j in range(len(sp)):

            if t[i] == sp[j]:

                tt[i][j] = 1

    return tt

def getData(fileName):

    df_train = pd.read_csv(fileName)

    y = convert(df_train['Species'])

    X = df_train.drop(['Id', 'Species'], axis=1).values

    return X, y

X_train, y_train = getData('iris_train.csv')

X_test, y_test = getData('iris_test.csv')

np.random.seed(1)

theta1 = 2 * np.random.rand(5, 4) - 1

theta2 = 2 * np.random.rand(5, 3) - 1

lamb = 0.1

for iter in range(1, 30001):

    # forwardPropagation

    m = len(X_train)

    a1 = np.hstack((np.ones([m, 1]), X_train))

    z2 = a1.dot(theta1)

    a2 = sigmoid(z2)

    a2 = np.hstack((np.ones([m, 1]), a2))

    z3 = a2.dot(theta2)

    a3 = sigmoid(z3)

    # backPropagation

    delta3 = (a3 - y_train).T

    if (iter % 10000) == 0:

        print("Error:" + str(np.mean(np.abs(delta3))))

    grad2 = np.dot(a2.T, delta3.T) + lamb * theta2

    grad2[0, :] -= lamb * theta2[0, :]

    delta2 = np.dot(theta2[1:, ], delta3) * sigmoidGradient(z2.T)

    delta2 = np.dot(theta2[1:, ], delta3) * (1 - a2.T[1:, ]) * a2.T[1:, ]

    grad1 = np.dot(a1.T, delta2.T) + lamb * theta1

    grad1[0, :] -= lamb * theta1[0, :]

    theta1 -= grad1 / m

    theta2 -= grad2 / m

res = np.argmax(predict(X_test, theta1, theta2), axis=1)

print(np.mean(np.argmax(y_test, axis = 1) == res))

以上就介绍了Python的相关知识，希望对Python有兴趣的朋友有所帮助。了解更多内容，请关注职坐标编程语言Python频道！