Question

我正在尝试添加函数calc_class生成的值，但它不起作用，我不知道原因。我试图使用numpy.append，numpy.insert和内置Python函数append失败。

这是我的代码：

def calc_class(test):
    expec = []
    for new in test:
        prob_vector = np.zeros((len(voc)), dtype=bool) #define a 'True' array to store class probabilities
        words_in_new = new[0].split() #split the new email into words
        words_in_new = list(set(words_in_new)) #remove duplicated words
        i = 0
        for voc_word in voc: #for each element in voc
            if voc_word in words_in_new:
                 prob_vector[i] = True #set the ith element of prob_vector to True, if voc element is in word
            else:
                prob_vector[i] = False #set the ith element of prob_vector to False, otherwise
            i += 1
        prob_ham = 1
        for i in range(len(prob_vector)):
            if prob_vector[i] == True:
                prob_ham *= ham_class_prob[i]
            else:
                prob_ham *= (1 - ham_class_prob[i])
        # alternative:     np.prod(ham_class_prob[np.where(prob_vector==True)]) * np.prod(1-  ham_class_prob[np.where(prob_vector==False)])

        prob_spam = 1
        for i in range(len(prob_vector)):
            if prob_vector[i] == True:
                prob_spam *= spam_class_prob[i]
            else:
                prob_spam *= (1 - spam_class_prob[i])

        p_spam = 0.3
        p_ham = 1 - p_spam

        p_spam_given_new = (prob_spam * p_spam) / (prob_spam * p_spam + prob_ham * p_ham)  # Bayes theorem
        print('p(spam|new_email)=', p_spam_given_new[0])
        expec.append(p_spam_given_new[0])
        print(expec)

问题是print(expect)正在打印一个空数组。

Answer 1

您可以使用pdb do debug（或ipdb for ipython）。

from pdb import set_trace

使用“set_trace（）”而不是“print（'p（spam | new_email）='，p_spam_given_new [0]）”（从行尾计算的第三行），而不是运行您的代码。它将在此行暂停，您可以在那里运行任何python代码，例如“print（p_spam_given_new）”或只是“p_spam_given_new”，您还可以检查“prob_spam”，“p_spam”或您要检查的任何其他变量。

无法将元素附加到numpy数组

1 个答案: