使用python进行性别分类模型的反向传播

Question

我在python中使用NLTK创建了一个性别分类模型。我正在使用前馈机制。如何将反向传播方法纳入其中？

import nltk
from nltk.corpus import names
import random

def gender_features(word):
    return {'last_letter': word[-1]}           

labeled_names=([(name1,'male') for name1 in names.words('male.txt')]+ 
           (name1,'female') for name1 in names.words('female.txt')])
random.shuffle(labeled_names)

featuresets = [(gender_features(n), gender) for (n, gender) in 
labeled_names]
train_set, test_set = featuresets[1500:], featuresets[:1500]
classifier = nltk.NaiveBayesClassifier.train(train_set)

#name input part
check=input('enter the name you wanna check:  ')
print(classifier.classify(gender_features(check)))
print(nltk.classify.accuracy(classifier, test_set))



feedback=input('is my answer correct??? [Y/N]:  ')
if feedback=='Y':
   print('thanks')
elif feedback=='N':
def gender_features1(word):
  return {'last_letter': word[-1],
          'second_last_letter': word[-2]}
labeled_names=([(name1,'male') for name1 in names.words('male.txt')]+
     [(name1,'female') for name1 in names.words('female.txt')])
  random.shuffle(labeled_names)

     featuresets = [(gender_features1(n), gender) for (n, gender) in labeled_names]
  train_set= featuresets[:1000]
  classifier = nltk.NaiveBayesClassifier.train(train_set)
  check1=input('plz enter the name again:  ')
  print(classifier.classify(gender_features(check1)))

 else:
     print('i dnt understand this input')

NaiveBayesClassifier接受feautresets，但我希望每次用户输入名称时使用新的输入数据重新训练分类器。

我还尝试过第二次使用两个功能集，但我不知道如何继续使用这个部分。