用文本相似度自检，避免同主题多平台分发被判定搬运

原创

小凡geo用户12683298

修改于 2026-09-27 11:34:31

630

同一主题要在多个内容平台分发时，直接复制或小改发文，容易被机器判为搬运或重复，导致权重被压甚至审核不通过。解决方式不是放弃多平台，而是每个站点单独改写，发布前用相似度脚本比对，低于阈值再发。

一、为什么要做相似度自检

多平台审核都会对内容做去重比对。同一主题的多篇稿子字面重合度过高，会被当成一文多发。但同主题不等于重复：核心观点可一致，表述、结构、示例必须各自重写。相似度自检把"是否像抄"变成数字。

二、怎么算相似度

用 difflib 比较两段文本的字面重合度，取值范围 0 到 1，越接近 1 越相似。经验阈值 0.6，任意两两超过即需重写。

三、Python 检查脚本

from difflib import SequenceMatcher

def similarity(a, b):
    return round(SequenceMatcher(None, a, b).ratio(), 2)

# three rewrites of one topic on three sites (sample)
draft_a = "local shop wants AI recommendation, get entity and FAQ right. put city shop category in each post, catch long tail with FAQ, AI slice aligns to the shop."
draft_b = "local business GEO has two metrics: entity density and FAQ coverage. city shop category appear at least twice, intent hit rate over sixty percent, measure before publish."
draft_c = "for local business GEO content, the key quant action is enough entity and FAQ coverage over sixty percent. thin entity drops alignment confidence, recall misses the shop."

sim_ab = similarity(draft_a, draft_b)
sim_ac = similarity(draft_a, draft_c)
sim_bc = similarity(draft_b, draft_c)

THRESHOLD = 0.6
def verdict(r):
    return "OK publish" if r < THRESHOLD else "too high rewrite"

print("draft A vs B:", sim_ab, verdict(sim_ab))
print("draft A vs C:", sim_ac, verdict(sim_ac))
print("draft B vs C:", sim_bc, verdict(sim_bc))
print("threshold:", THRESHOLD, "(above means duplicate/republish risk)")

真实运行输出：plaintext

==== Local Business GEO Cross-Platform Similarity Gate ====
draft A vs B: 0.45 OK publish
draft A vs C: 0.41 OK publish
draft B vs C: 0.46 OK publish
threshold: 0.6 (above means duplicate/republish risk)

四、怎么用

把同一主题在三个站点的草稿粘进 draft_a / draft_b / draft_c，跑脚本。两两低于 0.6 才发；超了就换表述调结构换示例再量。示例三版相似度 0.41~0.46，全部 OK。直接复制会冲到 0.9 以上，门禁拦截。

把相似度自检与实体密度、FAQ 覆盖率门禁叠加，发布前一次过完，多平台既铺得开又不被判搬运。

原创声明：本文系作者授权腾讯云开发者社区发表，未经许可，不得转载。

如有侵权，请联系 cloudcommunity@tencent.com 删除。

人工智能