应用支持向量机 (Support Vector Machine) 做垃圾邮件分类

Posted 2021-04-26 加载Python技能

学习思想：学习可以看作一个 输入-方法-输出 过程，敲几遍是一种有效的方式，是 Learn by Doing 的实践目标：应用线性支持向量机做垃圾邮件分类环境：python3, jupyter notebook, numpy, scipy, sklearn, nltk

算法

高斯核函数

def gaussian_kernel(x1, x2, sigma):
    x1 = x1.flatten()
    x2 = x2.flatten()
    sim = 0

    sim = np.exp(np.sum((x1-x2)**2)/(-2*sigma**2))
    return sim

处理邮件

def process_email(email_contents): 
    vocab_list = get_vocab_list()
    word_indices = np.array([], dtype=np.int64)

    email_contents = email_contents.lower()
    email_contents = re.sub('<[^<>]+>', ' ', email_contents)
    email_contents = re.sub('[0-9]+', 'number', email_contents)
    email_contents = re.sub('(http|https)://[^\s]*', 'httpaddr', email_contents)
    email_contents = re.sub('[^\s]+@[^\s]+', 'emailaddr', email_contents)
    email_contents = re.sub('[$]+', 'dollar', email_contents)
    print('==== Processed Email ====')

    stemmer = nltk.stem.porter.PorterStemmer()
    tokens = re.split('[@$/#.-:&*+=\[\]?!(){\},\'\">_<;% ]', email_contents)

    for token in tokens:
        token = re.sub('[^a-zA-Z0-9]', '', token)
        token = stemmer.stem(token)
        if len(token) < 1:
            continue
        for i in range(1, len(vocab_list) + 1):
            if vocab_list[i] == token:
                word_indices = np.append(word_indices, i)
        print(token)
    print('==================')
    return word_indices

def get_vocab_list():
    vocab_dict = {}
    with open('vocab.txt') as f:
        for line in f:
            (val, key) = line.split()
            vocab_dict[int(val)] = key
    return vocab_dict

电子邮件特征变量

def email_features(word_indices):
    n = 2018
    features = np.zeros(n+1)
    features[word_indices-1] = 1
    return features

具体步骤

步骤一，处理邮件

file_contents = open('emailSample1.txt', 'r').read()
word_indices = process_email(file_contents)

步骤二，特征变量

features = email_features(word_indices)

步骤三，训练线性支持向量机垃圾邮件分类

data = scio.loadmat('spamTrain.mat')
X = data['X']
y = data['y'].flatten()

c = 0.1
clf = svm.SVC(c, kernel='linear')
clf.fit(X, y)
p = clf.predict(X)

步骤四，测试垃圾分类

data = scio.loadmat('spamTest.mat')
Xtest = data['Xtest']
ytest = data['ytest'].flatten()
p = clf.predict(Xtest)

步骤五，打印最可能标示垃圾邮件的词汇

vocab_list = get_vocab_list()
indices = np.argsort(clf.coef_).flatten()[::-1]
print(indices)

for i in range(15):
    print('{} ({:0.6f})'.format(vocab_list[indices[i]], clf.coef_.flatten()[indices[i]]))