关于「知名度:评论数的对数边际、贝叶斯平均与近期速度」,《地图排名与同城承接》给出的定量结论是什么?

结论:知名度由三块组成。条数项是对数式的,下一条评论的边际增益从 0.03008 一路跌到 0.00010,老店堆评论抵不了半公里的距离。评分项对差评敏感,只有 10 条评论时,1 条一星要 63 条新评论才能消化。速度项是半衰期加权,突击换来的指数半年后只剩零头,稳定的来客才有稳定的指数。

数据口径:本书把三因子写成乘积式示意打分,全部权重都是示意参数,不是谷歌真实权重:一家距离两公里、类目登记错误的示意门店,要靠评论追平改对主类目的效果,需要 8,511 条评论。 书中参数为示意,全部计算由书中程序生成。

English version

对照数据表

下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.00010, 0.03008

nnext onenext tenoffset +0.5 kmoffset +1 km
50.030080.1913845771
200.007290.061072618,857
500.002670.0246490249,873
1000.001240.011882,343188,742
3000.000370.0036210,7591,583,338
1,0000.000100.0009757,56316,444,971
doubling a large n adds about a ln 2 = 0.2079 to P
查看完整验算程序
# Chapter 3. p31: the next review is worth less and less.
# Count term of the illustrative prominence: P(n) = 1 + a ln(1 + n),
# a = 0.30. Distance term exp(-d / 1.5 km). For a store with n reviews
# we print the gain from one more review, from ten more, and the
# review count that would offset a searcher 0.5 km or 1 km further
# away. Illustrative weights only, not Google's.
import math

A, D0 = 0.30, 1.5


def prom(n):
    return 1.0 + A * math.log1p(n)


def offset_count(n, extra_km):
    need = prom(n) * math.exp(extra_km / D0)
    return math.expm1((need - 1.0) / A)


if __name__ == "__main__":
    print("     n  next one   next ten   offset +0.5 km   offset +1 km")
    for n in (5, 20, 50, 100, 300, 1000):
        g1 = prom(n + 1) / prom(n) - 1.0
        g10 = prom(n + 10) / prom(n) - 1.0
        print("{:>6,}  {:.5f}   {:.5f}   {:>14,.0f}   {:>12,.0f}".format(
            n, g1, g10, offset_count(n, 0.5), offset_count(n, 1.0)))
    print("doubling a large n adds about a ln 2 = {:.4f} to P".format(
        A * math.log(2.0)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《本地搜索地图排名因子、评论速度信号权重与同城高急迫线索截流母典》第 3 章「知名度:评论数的对数边际、贝叶斯平均与近期速度」(版本 1.0.0). 5D Library. DOI 待发布

DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-369 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026369,
  author    = {Fan, Yafu},
  title     = {Local-Pack Search Ranking Factors, Review-Velocity Signalling and High-Urgency Geo-Intent Lead Capture},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论