Bigram字符预测模型

代码：

# 1.构建实验语料库
corpus = [
    "我喜欢吃苹果",
    "我喜欢吃香蕉",
    "她喜欢吃葡萄",
    "他不喜欢吃香蕉",
    "他喜欢吃苹果",
    "她喜欢吃草莓",
]

# 2.把句子分成N个“Gram”
def tokenize(text):
    return [char for char in text]

# 3.计算每个Bigram在语料库中的词频
from collections import defaultdict, Counter
def count_ngrams(corpus, n):
    ngrams_count = defaultdict(Counter)
    for text in corpus:
        tokens = tokenize(text)
        for i in range(len(tokens) - n + 1):
            ngram = tuple(tokens[i:i + n])
            prefix = ngram[:-1]
            token = ngram[-1]
            ngrams_count[prefix][token] += 1
    return ngrams_count

bigram_counts = count_ngrams(corpus,2)
# print("Bigram 词频：")
# for prefix, counts in bigram_counts.items():
#     print("{}:{}".format("".join(prefix), dict(counts)))



# 4.计算每个Bigram出现的频率
def ngram_probabilities(ngrams_coun