ARPA的n-gram語法如下：

[html] view plaincopyprint? \data\ ngram 1=64000 ngram 2=522530 ngram 3=173445 \1-grams: -5.24036 'cause -0.2084827 -4.675221 'em -0.221857 -4.989297 'n -0.05809768 -5.365303 'til -0.1855581 -2.111539 </s> 0.0 -99 <s> -0.7736475 -1.128404 <unk> -0.8049794 -2.271447 a -0.6163939 -5.174762 a's -0.03869072 -3.384722 a. -0.1877073 -5.789208 a.'s 0.0 -6.000091 aachen 0.0 -4.707208 aaron -0.2046838 -5.580914 aaron's -0.06230035 -5.789208 aarons -0.07077657 -5.881973 aaronson -0.2173971

具體說明見 ：ARPA的n-gram語言模型格式

整個ARPA-LM由很多個n-gram項組成，分別說明這兩個的數據結構

一，n-gram數據結構

n-gram的數據結構如下：

typedef struct { real log_prob ; real log_bo ; int *words ; } ARPALMEntry ;

words，表示當前的n-gram所涉及的單詞，如果是1-gram，那就只有一個，如果是2-gram，那么words就包括這兩個單詞的序號。
log_bo，表示ngram的回退概率。
log_prob，表示ngram的組合概率。

二，ARPA-LM數據結構

多個項組成的整個n-gram語言模型的數據結構如下：
[cpp] view plaincopyprint?

class ARPALM { public: Vocabulary *vocab ; int order ; ARPALMEntry **entries ; // 語言模型的所有項，組成一個數組 int *n_ngrams ; // 一元語言模型、二元語言模型、三元語言模型等組成的數組，數組每一項都表示對應的的元有多少個。 char *unk_wrd ; // 詞典中不在語言模型中的詞。 int unk_id ;// 詞典中不在語言模型中的詞ID，這個ID指定為詞典的最后一個序號。 int n_unk_words ; int *unk_words ; private: bool *words_in_lm ; // 布爾類型數組，標識詞是否在語言模型中。 }

vocab，用于構建語言模型的詞典指針。詞典定義見：詞典內存存儲模型
entries，語言模型的所有ngram項，是ARPALMEntry類型的一個二維數組。entries[0]存儲1-gram，entries[1]存儲2-gram，依此類推。
n_ngrams，整型數組，依次包含1-gram,2-gram,3-gram,....所包含的ngram項個數。
unk_wrd，詞典中可以不在語言模型中的詞。
unk_id，詞典中可以不在語言模型中的詞的ID，這個ID指定為詞典的最后一個詞序號。
n_unk_words，在讀語言模型之后，統計在詞典中，但沒有用來建立語言模型的詞個數，如果沒有指定unk_wrd的話，是不允許的，就表示所有的詞典中的詞都應該用來建語言模型。
unk_words，存儲6中統計的詞序號。
words_in_lm，這個標識詞典中的詞是否在語言模型中出現。

轉載于:https://www.cnblogs.com/jonky/p/10154115.html

總結

以上是生活随笔為你收集整理的N-Gram的数据结构的全部內容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網站內容還不錯，歡迎將生活随笔推薦給好友。