关于「原典考辨:郑樵《通志·校雠略》的类例与布什的联想路径」,《长尾页信息增益与收录排程》给出的定量结论是什么?

结论:第七章考辨郑樵《通志·校雠略》的类例与布什 1945 年的联想路径,并给出四件套蒸馏规格。

数据口径:书中参数为示意,全部计算由书中程序生成。 本书是非个人化经营建议;房产内容不构成法律、税务或投资建议,塔罗部分不做个人占断,宗祠资料须经本人同意,书中数据全部合成。

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

A. leaves N, links per category page b
Nblevelscategory pagesclicks home->leaf
39383583
393202212
39360282
393400111
2,880844124
2,8802031533
2,880602492
2,880400292
B. coherence of 12 categories over 393 leaves
grouped by area0.5854
grouped at random0.1259
查看完整验算程序
# Chapter 7. p71: Zheng Qiao's lei li (classes and their order) as a
# site hierarchy. A. For N leaf pages and at most b links per category
# page, the shallowest tree needs ceil(log_b N) levels of categories;
# we count category pages and clicks from home. B. "Class by subject":
# group the 393 kept property pages of p21-style data by area, or at
# random, into the same number of categories, and measure coherence as
# the share of sibling pairs that share area and property type. Data
# are synthetic (fixed seed). The four-kit spec of this chapter keeps
# breadcrumbs within 3 levels and holds the front-end floors of the
# house style library: tap targets >= 44 px, text contrast >= 4.5:1.
import itertools
import math
import random

rng = random.Random(20261120)


def levels(n, b):
    lv, cap = 0, 1
    while cap < n:
        cap *= b
        lv += 1
    return lv


def cat_pages(n, b):
    total, width = 0, n
    while width > 1:
        width = int(math.ceil(width / float(b)))
        total += width
    return total


def coherence(groups):
    same, pairs = 0, 0
    for g in groups:
        for x, y in itertools.combinations(g, 2):
            pairs += 1
            same += (x[0] == y[0]) + (x[1] == y[1])
    return same / (2.0 * pairs)


if __name__ == "__main__":
    print("A. leaves N, links per category page b")
    print("      N     b  levels  category pages  clicks home->leaf")
    for n in (393, 2880):
        for b in (8, 20, 60, 400):
            lv = levels(n, b)
            fmt = "  {:>5,}  {:>4}    {:>2}       {:>5,}            {}"
            print(fmt.format(n, b, lv, cat_pages(n, b), lv))
    leaves = [(rng.randrange(12), rng.randrange(6), rng.randrange(8))
              for _ in range(393)]
    by_area = [[x for x in leaves if x[0] == a] for a in range(12)]
    shuffled = leaves[:]
    rng.shuffle(shuffled)
    sizes = [len(g) for g in by_area]
    rand, pos = [], 0
    for s in sizes:
        rand.append(shuffled[pos:pos + s])
        pos += s
    print("B. coherence of 12 categories over 393 leaves")
    print("  grouped by area      {:.4f}".format(coherence(by_area)))
    print("  grouped at random    {:.4f}".format(coherence(rand)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《多域名程序化SEO长尾拓扑、新站沙盒极速收录与高意图自然流量收割母典》第 7 章「原典考辨:郑樵《通志·校雠略》的类例与布什的联想路径」(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-346 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026346,
  author    = {Fan, Yafu},
  title     = {Programmatic SEO Long-Tail Topology, Rapid Indexation and High-Intent Organic Traffic Harvesting across Multi-Domain Fleets},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论