按类别分组的句子中最常用的词

Question

我正在尝试按类别对 10 个最常用的单词进行分组。我已经看到了这个答案，但我不能完全修改它以获得我想要的输出。category | sentence  A           cat runs over big dog  A           dog runs over big cat  B           random sentences include words  C           including this one所需的输出：category | word/frequency   A           runs, 2               cat: 2               dog: 2               over: 2               big: 2   B           random: 1   C           including: 1由于我的数据框非常大，我只想获得前 10 个最常出现的词。我也看过这个答案df.groupby('subreddit').agg(lambda x: nltk.FreqDist([w for wordlist in x for w in wordlist]))但此方法也返回字母数。

Answer 1

# Split the sentence into Series    

df1 = pd.DataFrame(df.sentence.str.split(' ').tolist())

# Add category with as not been adding with the split

df1['category']  = df['category']

# Melt the Series corresponding to the splited sentence

df1 = pd.melt(df1, id_vars='category', value_vars=df1.columns[:-1].tolist())

# Groupby and count (reset_index will create a column nammed 0)

df1 = df1.groupby(['category', 'value']).size().reset_index()

# Keep the 10 largests numbers 

df1 = df1.nlargest(10, 0)