def __tfidf_vectors(values): tfidf_matrix = TfidfVectorizer().fit_transform(values) split_index = int(tfidf_matrix.get_shape()[0] / 2) tfidf_array = tfidf_matrix.todense() df_result1 = [row.tolist()[0] for row in tfidf_array[0:split_index]] df_result2 = [row.tolist()[0] for row in tfidf_array[split_index:]] return df_result1, df_result2