\u8fc7\u53bb\u4e24\u5e74，AI\u4e0e\u4f20\u7edf\u6587\u5316\u7684\u7ed3\u5408\u65b9\u5f0f\u53ef\u4ee5\u6982\u62ec\u4e3a"\u5a31\u4e50\u5316"\u2014\u2014AI\u5199\u53e4\u8bd7\u3001AI\u753b\u6c34\u58a8\u753b\u3001AI\u751f\u6210\u5bf9\u8054\u3002\u8fd9\u4e9b\u5e94\u7528\u4f20\u64ad\u5e7f\u3001\u8da3\u5473\u6027\u5f3a，\u4f46\u672c\u8d28\u4e0a\u662f\u7528AI\u7684\u751f\u6210\u80fd\u529b\u4e3a\u4f20\u7edf\u6587\u5316"\u62ab\u4e0a\u4e86\u4e00\u5c42\u79d1\u6280\u5916\u8863"\u3002<\/p>\n
\u8fd9\u4e0eAI\u5728\u5176\u4ed6\u9886\u57df\u7684\u5e94\u7528\u5f62\u6210\u4e86\u9c9c\u660e\u5bf9\u6bd4\u3002\u5f53AI\u5728\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u4e0a\u62ff\u8bfa\u8d1d\u5c14\u5956\u65f6，\u5b83\u5728\u4f20\u7edf\u6587\u5316\u9886\u57df\u7684\u5e94\u7528\u8fd8\u505c\u7559\u5728"\u597d\u73a9"\u7684\u5c42\u9762\u3002\u4e0d\u662f\u4f20\u7edf\u6587\u5316\u6ca1\u6709\u6df1\u5ea6，\u800c\u662f\u6211\u4eec\u5bf9AI\u7684\u7528\u6cd5\u8fd8\u4e0d\u591f"\u7814\u7a76\u65b9\u6cd5\u8bba\u5316"\u3002<\/p>\n
\u89c1\u8bc1\u5947\u8ff9\u7684\u65f6\u523b，\u4e0d\u5728\u4e8eAI\u753b\u51fa\u4e00\u5e45\u4ee5\u5047\u4e71\u771f\u7684\u6c34\u58a8\u753b，\u800c\u5728\u4e8eAI\u5e2e\u52a9\u7814\u7a76\u8005\u53d1\u73b0\u300a\u8bd7\u7ecf\u300b\u4e2d\u9690\u85cf\u7684\u97f3\u97f5\u89c4\u5f8b\u2014\u2014\u8fd9\u79cd\u53d1\u73b0\u662f\u4f20\u7edf\u4eba\u5de5\u7814\u7a76\u53ef\u80fd\u9700\u8981\u6570\u5e74\u624d\u80fd\u5b8c\u6210\u7684\u3002\u4e0b\u534a\u573a\u7684\u5173\u952e\u8f6c\u53d8\u662f：\u7528AI\u505a"\u7814\u7a76"，\u800c\u4e0d\u662f\u7528AI\u505a"\u8868\u6f14"\u3002<\/p>\n
AI\u4e0e\u4f20\u7edf\u6587\u5316\u878d\u5408\u7684\u65b9\u6cd5\u8bba\u5347\u7ea7\u8def\u5f84：<\/p>\n
\u4e0a\u534a\u573a\u7684\u6a21\u5f0f："\u7528AI\u505a\u4f20\u7edf\u6587\u5316\u5185\u5bb9"\u3002 \u672c\u8d28\u662f\u5185\u5bb9\u751f\u4ea7\u5de5\u5177\u7684\u5347\u7ea7\u2014\u2014\u5c06AI\u5f53\u4f5c\u4e00\u4e2a\u66f4\u9ad8\u6548\u7684"\u521b\u4f5c\u5de5\u5177"\u3002\u95ee\u9898\u5728\u4e8e：\u8fd9\u4e9b\u5e94\u7528\u505c\u7559\u5728\u4f20\u7edf\u6587\u5316\u7684\u8868\u9762\u5f62\u5f0f，\u6ca1\u6709\u6df1\u5165\u7406\u89e3\u5176\u5185\u5728\u903b\u8f91\u3002<\/p>\n
\u4e0b\u534a\u573a\u7684\u6a21\u5f0f： "\u7528AI\u505a\u4f20\u7edf\u6587\u5316\u7814\u7a76"\u3002\u8fd9\u662f\u65b9\u6cd5\u8bba\u7684\u5347\u7ea7\u3002\u5de5\u5177\u8fd8\u662f\u90a3\u4e9b\u5de5\u5177（NLP\u3001\u77e5\u8bc6\u56fe\u8c31\u3001\u7edf\u8ba1\u5206\u6790），\u4f46\u4f7f\u7528\u65b9\u5f0f\u53d8\u4e86：<\/p>\n
\u5173\u952e\u8f6c\u53d8\u3002 \u4ece"AI\u80fd\u751f\u6210\u4ec0\u4e48"\u8f6c\u5411"AI\u80fd\u53d1\u73b0\u4ec0\u4e48"\u3002<\/p>\n
\u7528NLP\u65b9\u6cd5\u505a\u5927\u89c4\u6a2a\u53e4\u7c4d\u6587\u672c\u5206\u6790：<\/p>\n
import re
\nfrom typing import List, Dict, Tuple
\nfrom collections import Counter, defaultdict
\nimport math<\/p>\n
class AncientTextAnalyzer:
\n """
\n \u53e4\u7c4d\u6587\u672c\u91cf\u5316\u5206\u6790\u5668<\/p>\n
\u8bbe\u8ba1\u539f\u56e0：\u4f20\u7edf\u4eba\u6587\u7814\u7a76\u4f9d\u8d56"\u7ec6\u8bfb"(close reading)，
\n \u7814\u7a76\u8005\u53ea\u80fd\u5206\u6790\u6709\u9650\u6587\u672c\u3002\u8ba1\u7b97\u5206\u6790\u652f\u6301"\u8fdc\u8bfb"(distant reading)，
\n \u5728\u4e07\u5377\u7ea7\u522b\u53d1\u73b0\u5b8f\u89c2\u6a21\u5f0f\u3002\u4e24\u79cd\u65b9\u6cd5\u4e92\u8865\u800c\u975e\u66ff\u4ee3\u3002
\n """<\/p>\n
def __init__(self, stop_words_file: str = None):
\n # \u8bbe\u8ba1\u539f\u56e0：\u53e4\u6c49\u8bed\u505c\u7528\u8bcd\u4e0e\u73b0\u4ee3\u6c49\u8bed\u4e0d\u540c，
\n # \u9700\u52a0\u8f7d\u4e13\u95e8\u7684\u505c\u7528\u8bcd\u8868\u4ee5\u907f\u514d\u8fc7\u6ee4\u6389\u6709\u4fe1\u606f\u91cf\u7684\u865a\u8bcd
\n self.stop_words = set()
\n if stop_words_file:
\n with open(stop_words_file, 'r', encoding='utf-8') as f:
\n self.stop_words = set(line.strip() for line in f)<\/p>\n
def segment_poetry(self, text: str) -> List[str]:
\n """
\n \u8bd7\u6b4c\u5206\u8bcd<\/p>\n
\u8bbe\u8ba1\u539f\u56e0：\u53e4\u6c49\u8bed\u5355\u5b57\u6210\u8bcd\u7684\u6bd4\u4f8b\u8fdc\u9ad8\u4e8e\u73b0\u4ee3\u6c49\u8bed，
\n \u5206\u8bcd\u7b56\u7565\u5e94\u504f\u5411\u7ec6\u7c92\u5ea6（\u6309\u5b57\u5207\u5206+\u57fa\u672c\u8bcd\u8bed\u5408\u5e76），
\n \u800c\u975e\u7167\u642c\u73b0\u4ee3\u6c49\u8bed\u7684\u5206\u8bcd\u7c92\u5ea6
\n """
\n # \u8bbe\u8ba1\u539f\u56e0：\u5148\u6309\u6807\u70b9\u548c\u6362\u884c\u5207\u53e5，
\n # \u8bd7\u8bcd\u7684\u53e5\u95f4\u505c\u987f\u6709\u660e\u786e\u7684\u4fe1\u606f\u8fb9\u754c
\n sentences = re.split(r'[，\u3002！？；\u3001\\\\n]+', text)
\n tokens = []
\n for sentence in sentences:
\n # \u8bbe\u8ba1\u539f\u56e0：\u9010\u5b57\u5207\u5206\u4fdd\u7559\u6700\u5927\u4fe1\u606f\u91cf，
\n # \u53e4\u6c49\u8bed\u7814\u7a76\u8005\u66f4\u4e60\u60ef\u5b57\u7ea7\u522b\u7684\u5206\u6790
\n chars = list(sentence.strip())
\n tokens.extend(c for c in chars if c.strip())
\n return tokens<\/p>\n
def analyze_rhyme_pattern(
\n self, poems: List[str]
\n ) -> Dict[str, float]:
\n """
\n \u97f5\u811a\u6a21\u5f0f\u5206\u6790<\/p>\n
\u8bbe\u8ba1\u539f\u56e0：\u97f5\u811a\u662f\u53e4\u5178\u8bd7\u6b4c\u7684\u6838\u5fc3\u5f62\u5f0f\u7279\u5f81\u3002
\n \u81ea\u52a8\u5206\u6790\u97f5\u811a\u5206\u5e03\u53ef\u4ee5\u91cf\u5316\u4e0d\u540c\u65f6\u671f\u3001\u4e0d\u540c\u6d41\u6d3e\u7684\u7528\u97f5\u504f\u597d\u3002
\n """
\n rhyme_counter = Counter()
\n total_lines = 0<\/p>\n
for poem in poems:
\n lines = [l.strip() for l in poem.split('\\\\n') if l.strip()]
\n for line in lines:
\n if line:
\n last_char = line[-1]
\n rhyme_counter[last_char] += 1
\n total_lines += 1<\/p>\n
# \u8bbe\u8ba1\u539f\u56e0：\u5f52\u4e00\u5316\u4e3a\u9891\u7387\u800c\u975e\u7edd\u5bf9\u8ba1\u6570，
\n # \u6d88\u9664\u4e0d\u540c\u6837\u672c\u91cf\u5e26\u6765\u7684\u504f\u5dee
\n return {
\n char: count \/ total_lines
\n for char, count in rhyme_counter.most_common(50)
\n }<\/p>\n
def compute_tf_idf_across_works(
\n self, works: Dict[str, str]
\n ) -> Dict[str, List[Tuple[str, float]]]:
\n """
\n \u8de8\u4f5c\u54c1\u7684TF-IDF\u5206\u6790<\/p>\n
\u8bbe\u8ba1\u539f\u56e0：TF-IDF\u53ef\u4ee5\u627e\u51fa\u6bcf\u90e8\u4f5c\u54c1\u72ec\u6709\u7684\u5173\u952e\u8bcd，
\n \u8fd9\u4e9b\u8bcd\u5f80\u5f80\u53cd\u6620\u4e86\u4f5c\u54c1\u7684\u72ec\u7279\u4e3b\u9898\u548c\u98ce\u683c\u3002
\n \u4f8b\u5982：\u675c\u752b\u7684\u8bd7\u53ef\u80fd\u9ad8\u9891\u51fa\u73b0"\u60b2"\u3001"\u56fd"\u7b49\u8bcd，
\n \u800c\u674e\u767d\u7684\u8bd7\u53ef\u80fd\u9ad8\u9891\u51fa\u73b0"\u9152"\u3001"\u6708"\u3002
\n """
\n # \u6587\u6863\u603b\u6570
\n N = len(works)
\n doc_tokens = {}
\n df = Counter() # document frequency<\/p>\n
# \u5206\u8bcd\u548c\u6587\u6863\u9891\u7387\u7edf\u8ba1
\n for work_name, text in works.items():
\n tokens = self.segment_poetry(text)
\n unique_tokens = set(tokens)
\n doc_tokens[work_name] = tokens
\n for token in unique_tokens:
\n df[token] += 1<\/p>\n
# \u8ba1\u7b97\u6bcf\u90e8\u4f5c\u54c1\u7684TF-IDF
\n results = {}
\n for work_name, tokens in doc_tokens.items():
\n tf = Counter(tokens)
\n total = len(tokens)<\/p>\n
tfidf_scores = {}
\n for term, freq in tf.items():
\n if term in self.stop_words:
\n continue
\n # \u8bbe\u8ba1\u539f\u56e0：\u5e73\u6ed1\u5904\u7406\u9632\u6b62df=0\u5bfc\u81f4\u9664\u96f6
\n idf = math.log((N + 1) \/ (df[term] + 1)) + 1
\n tfidf_scores[term] = (freq \/ total) * idf<\/p>\n
results[work_name] = sorted(
\n tfidf_scores.items(),
\n key=lambda x: x[1],
\n reverse=True
\n )[:20]<\/p>\n
return results<\/p>\n
def detect_thematic_shifts(
\n self, periods: Dict[str, List[str]]
\n ) -> Dict[str, str]:
\n """
\n \u4e3b\u9898\u53d8\u8fc1\u68c0\u6d4b<\/p>\n
\u8bbe\u8ba1\u539f\u56e0：\u901a\u8fc7\u5bf9\u6bd4\u4e0d\u540c\u65f6\u671f\u7684\u8bcd\u9891\u5206\u5e03，
\n \u91cf\u5316\u5de5\u53d1\u73b0\u6587\u5316\u5173\u6ce8\u70b9\u7684\u53d8\u8fc1\u3002
\n \u5982\u5510\u4ee3\u8bd7\u6b4c\u591a\u7528"\u8fb9\u585e"\u3001"\u80e1"\u7b49\u8bcd，
\n \u5b8b\u4ee3\u8bd7\u6b4c\u591a\u7528"\u4e66\u658b"\u3001"\u8336"\u7b49\u8bcd\u3002
\n """
\n period_word_freq = {}<\/p>\n
for period, texts in periods.items():
\n all_tokens = []
\n for text in texts:
\n all_tokens.extend(self.segment_poetry(text))
\n top_words = Counter(all_tokens).most_common(30)
\n period_word_freq[period] = dict(top_words)<\/p>\n
return period_word_freq<\/p>\n
\u8ba1\u7b97\u5206\u6790 vs \u4f20\u7edf\u7ec6\u8bfb\u3002 \u8ba1\u7b97\u5206\u6790\u64c5\u957f\u53d1\u73b0\u5b8f\u89c2\u6a21\u5f0f（"\u5510\u8bd7\u4e2d\u81ea\u7136\u610f\u8c61\u7684\u4f7f\u7528\u9891\u7387\u662f\u5b8b\u8bd7\u76841.7\u500d"），\u4f46\u65e0\u6cd5\u6355\u6349\u6587\u672c\u7684\u5fae\u5999\u4e4b\u5904（"\u8fd9\u53e5\u8bd7\u7684\u610f\u5883\u7f8e\u5728\u54ea\u91cc"）\u3002\u4f20\u7edf\u7ec6\u8bfb\u80fd\u6df1\u5165\u7406\u89e3\u5355\u4e2a\u6587\u672c，\u4f46\u65e0\u6cd5\u5904\u7406\u5927\u89c4\u6a21\u8bed\u6599\u3002\u4e24\u79cd\u65b9\u6cd5\u7ed3\u5408：\u8ba1\u7b97\u53d1\u73b0\u5047\u8bbe，\u7ec6\u8bfb\u9a8c\u8bc1\u5047\u8bbe\u3002<\/p>\n
\u5ba2\u89c2\u91cf\u5316 vs \u4e3b\u89c2\u8be0\u91ca\u3002 \u8ba1\u7b97\u5206\u6790\u7684\u8f93\u51fa\u662f\u6570\u5b57\u2014\u2014\u9891\u7387\u3001\u5206\u5e03\u3001\u76f8\u5173\u6027\u3002\u8fd9\u4e9b\u6570\u5b57\u672c\u8eab\u6ca1\u6709\u610f\u4e49，\u9700\u8981\u4eba\u6587\u7814\u7a76\u8005\u505a\u51fa\u8be0\u91ca\u3002\u6570\u5b57\u53ea\u80fd\u8bf4"\u662f\u4ec0\u4e48"，\u4e0d\u80fd\u8bf4"\u4e3a\u4ec0\u4e48"\u3002\u8fd9\u4e5f\u662f\u4e3a\u4ec0\u4e48AI\u5de5\u5177\u4e0d\u4f1a\u53d6\u4ee3\u4eba\u6587\u7814\u7a76\u8005，\u800c\u662f\u7ed9\u4ed6\u4eec\u66f4\u5f3a\u5927\u7684"\u53d1\u73b0\u5de5\u5177"\u3002<\/p>\n
\u901a\u7528NLP\u5de5\u5177 vs \u53e4\u6c49\u8bed\u4e13\u7528\u5de5\u5177\u3002 \u901a\u7528\u5206\u8bcd\u5668\u5bf9\u73b0\u4ee3\u6c49\u8bed\u6709\u6548，\u5bf9\u53e4\u6c49\u8bed\u6548\u679c\u5f88\u5dee\u3002\u76f4\u63a5\u7528BGE\u7b49\u901a\u7528\u5d4c\u5165\u6a21\u578b\u5904\u7406\u53e4\u6587，\u8bed\u4e49\u7a7a\u95f4\u4f1a\u53d1\u751f\u504f\u79fb\u3002\u9700\u8981\u53e4\u6c49\u8bed\u4e13\u7528\u7684\u5206\u8bcd\u5668\u548c\u5d4c\u5165\u6a21\u578b\u2014\u2014\u4f46\u6784\u5efa\u8fd9\u4e9b\u5de5\u5177\u7684\u6210\u672c\u5f88\u9ad8，\u53d7\u4f17\u53c8\u6709\u9650\u3002<\/p>\n
\u5e7f\u5ea6 vs \u6df1\u5ea6\u3002 \u8ba1\u7b97\u5206\u6790\u7684\u4f18\u52bf\u662f\u5e7f\u5ea6\u2014\u2014\u53ef\u4ee5\u5904\u7406\u6570\u5343\u90e8\u4f5c\u54c1\u3002\u52a3\u52bf\u662f\u6df1\u5ea6\u2014\u2014\u6bcf\u6761\u5206\u6790\u90fd\u76f8\u5bf9\u6d45\u5c42\u3002\u6298\u4e2d\u65b9\u6848：\u7528\u8ba1\u7b97\u5206\u6790\u505a\u5168\u5c40\u626b\u63cf\u548c\u5047\u8bbe\u751f\u6210，\u5bf9\u53d1\u73b0\u7684\u5173\u952e\u6a21\u5f0f\u7528\u4f20\u7edf\u65b9\u6cd5\u505a\u6df1\u5ea6\u7814\u7a76\u3002<\/p>\n
AI\u4e0e\u4f20\u7edf\u6587\u5316\u878d\u5408\u6b63\u5728\u4ece"\u5a31\u4e50\u5316\u5e94\u7528"\u9636\u6bb5\u8fdb\u5165"\u7814\u7a76\u65b9\u6cd5\u8bba"\u9636\u6bb5\u3002\u4e0a\u534a\u573a\u7684\u95ee\u9898\u662f：AI\u88ab\u5f53\u4f5c\u5185\u5bb9\u751f\u6210\u5de5\u5177，\u4ea7\u51fa\u4e86\u5927\u91cf"\u50cf\u90a3\u4e48\u56de\u4e8b"\u4f46\u7f3a\u4e4f\u7814\u7a76\u4ef7\u503c\u7684\u5185\u5bb9\u3002\u4e0b\u534a\u573a\u7684\u65b9\u5411\u662f：\u5c06NLP\u3001\u77e5\u8bc6\u56fe\u8c31\u3001\u7edf\u8ba1\u5206\u6790\u7b49AI\u6280\u672f\u4f5c\u4e3a\u7814\u7a76\u5de5\u5177，\u5e2e\u52a9\u4eba\u6587\u7814\u7a76\u8005\u5728\u5927\u89c4\u6a21\u6587\u672c\u4e2d\u53d1\u73b0\u4e4b\u524d\u96be\u4ee5\u53d1\u73b0\u7684\u5b8f\u89c2\u6a21\u5f0f\u3002\u8fd9\u9700\u8981\u7684\u4e0d\u662f\u66f4\u5f3a\u7684\u751f\u6210\u6a21\u578b，\u800c\u662f\u66f4\u4e25\u8c28\u7684\u7814\u7a76\u65b9\u6cd5\u8bba\u2014\u2014\u8ba1\u7b97\u4eba\u6587\u3001\u8fdc\u8bfb\u65b9\u6cd5\u3001\u6570\u5b57\u4eba\u6587\u7b49\u5b66\u79d1\u6b63\u5728\u4e3a\u8fd9\u4e2a\u65b9\u5411\u63d0\u4f9b\u7406\u8bba\u6846\u67b6\u3002<\/p>\n
\u672c\u6587\u4e2d\u7684\u534f\u8bae\u3001\u7248\u672c\u3001\u6027\u80fd\u3001\u6210\u672c\u548c\u884c\u4e1a\u8d8b\u52bf\u5e94\u4ee5\u53ef\u6838\u9a8c\u7684\u4e00\u624b\u8d44\u6599\u4e3a\u51c6\u3002\u672a\u6807\u6ce8\u7edf\u8ba1\u53e3\u5f84\u7684\u6bd4\u4f8b\u3001\u65f6\u95f4\u8868\u548c\u9884\u6d4b\u4ec5\u4f5c\u5de5\u7a0b\u8ba8\u8bba，\u4e0d\u5e94\u89c6\u4e3a\u884c\u4e1a\u4e8b\u5b9e\u3002\u53ef\u53c2\u8003 0730 \u8d44\u6599\u6765\u6e90\u7d22\u5f15，\u5e76\u5728\u53d1\u5e03\u524d\u5c06\u5177\u4f53\u6765\u6e90\u8d34\u5230\u5bf9\u5e94\u65ad\u8a00\u4e4b\u540e\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"
AI \u4e0e\u4f20\u7edf\u6587\u5316\u878d\u5408\u7684\u4e0b\u534a\u573a：\u4ece\u5a31\u4e50\u5230\u7814\u7a76\u7684\u65b9\u6cd5\u8bba\u5347\u7ea7
\n\u4e00\u3001\u4e2a\u6027\u5316\u6df1\u5ea6\u5f15\u8a00
\n\u8fc7\u53bb\u4e24\u5e74，AI\u4e0e\u4f20\u7edf\u6587\u5316\u7684\u7ed3\u5408\u65b9\u5f0f\u53ef\u4ee5\u6982\u62ec\u4e3a\\”\u5a31\u4e50\u5316\\”\u2014\u2014AI\u5199\u53e4\u8bd7\u3001AI\u753b\u6c34\u58a8\u753b\u3001AI\u751f\u6210\u5bf9\u8054\u3002\u8fd9\u4e9b\u5e94\u7528\u4f20\u64ad\u5e7f\u3001\u8da3\u5473\u6027\u5f3a，\u4f46\u672c\u8d28\u4e0a\u662f\u7528AI\u7684\u751f\u6210\u80fd\u529b\u4e3a\u4f20\u7edf\u6587\u5316\\”\u62ab\u4e0a\u4e86\u4e00\u5c42\u79d1\u6280\u5916\u8863\\”\u3002
\n\u8fd9\u4e0eAI\u5728\u5176\u4ed6\u9886\u57df\u7684\u5e94\u7528\u5f62\u6210\u4e86\u9c9c\u660e\u5bf9\u6bd4\u3002\u5f53AI\u5728\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u4e0a\u62ff\u8bfa\u8d1d\u5c14\u5956\u65f6，\u5b83\u5728\u4f20\u7edf\u6587\u5316\u9886\u57df\u7684\u5e94\u7528\u8fd8\u505c\u7559\u5728\\”\u597d\u73a9\\”\u7684\u5c42\u9762\u3002\u4e0d\u662f\u4f20\u7edf\u6587\u5316\u6ca1\u6709\u6df1\u5ea6࿰<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[50],"topic":[],"class_list":["post-88909","post","type-post","status-publish","format-standard","hentry","category-server","tag-50"],"yoast_head":"\n