关于「关键词矩阵:笛卡尔积、可行组合与先建哪一百页」,《长尾页信息增益与收录排程》给出的定量结论是什么?

结论:第二章用三道剪刀把 2,880 个关键词格子剪到 393 页。

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 标准库与结论吻合的打印值: 2880, 393

property matrix 12 x 6 x 5 x 8
raw Cartesian cells2,880
feasible combinations2,352
demand >10 a month 1,022
and >8 own facts 393
demand kept 19,538 of 49,3450.3959
pages kept / raw cells0.1365
tarot matrix 78 x 2 x 4624 cells
distinct meaning texts available 234
cells per distinct text 2.67
查看完整验算程序
# Chapter 2. p21: how a Cartesian keyword matrix shrinks under honest
# pruning. Matrix for a property site: 12 areas x 6 property types x
# 5 price bands x 8 buyer questions. Rules (illustrative): some types do
# not exist in some areas or bands; a cell needs >= 10 expected searches
# a month and >= 8 distinct facts from the site's own data (listings,
# transactions, by-law notes) to earn a page. Demand per cell is a
# product of Zipf weights per dimension; data counts are drawn with a
# fixed seed. A second matrix is the tarot site: 78 cards x 2
# orientations x 4 page types, where most cells share one meaning text.
import itertools
import math
import random

rng = random.Random(20261120)
AREAS, TYPES, BANDS, QUESTIONS = 12, 6, 5, 8
V_TOTAL = 60000.0          # searches a month across the whole matrix
NO_TYPE = {(0, 4), (0, 5), (1, 5), (2, 5), (3, 4)}   # (area, type)
NO_BAND = {(4, 0), (5, 0), (4, 1), (0, 4)}          # (type, band)


def zipf_w(n, s=1.0):
    h = math.fsum(r ** -s for r in range(1, n + 1))
    return [r ** -s / h for r in range(1, n + 1)]


WA, WT = zipf_w(AREAS), zipf_w(TYPES, 0.8)
WB, WQ = zipf_w(BANDS, 0.6), zipf_w(QUESTIONS, 1.1)


def cells():
    for a, t, b, q in itertools.product(range(AREAS), range(TYPES),
                                        range(BANDS), range(QUESTIONS)):
        demand = V_TOTAL * WA[a] * WT[t] * WB[b] * WQ[q]
        facts = rng.randint(0, 6) + rng.randint(0, 6) * (a < 6) + \
            rng.randint(0, 4) * (q < 4)
        yield a, t, b, q, demand, facts


if __name__ == "__main__":
    rows = list(cells())
    raw = len(rows)
    feas = [r for r in rows if (r[0], r[1]) not in NO_TYPE
            and (r[1], r[2]) not in NO_BAND]
    dem = [r for r in feas if r[4] >= 10.0]
    both = [r for r in dem if r[5] >= 8]
    print("property matrix 12 x 6 x 5 x 8")
    print("  raw Cartesian cells       {:>6,}".format(raw))
    print("  feasible combinations     {:>6,}".format(len(feas)))
    print("  demand >= 10 a month      {:>6,}".format(len(dem)))
    print("  and >= 8 own facts        {:>6,}".format(len(both)))
    fd = math.fsum(r[4] for r in feas)
    kd = math.fsum(r[4] for r in both)
    print("  demand kept {:,.0f} of {:,.0f} = {:.4f}".format(kd, fd, kd / fd))
    print("  pages kept / raw cells = {:.4f}".format(len(both) / raw))
    tarot = 78 * 2 * 4
    distinct = 78 * 2 + 78
    print("tarot matrix 78 x 2 x 4 = {:,} cells".format(tarot))
    print("  distinct meaning texts available {:,}".format(distinct))
    print("  cells per distinct text {:.2f}".format(tarot / distinct))

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 2 章「关键词矩阵:笛卡尔积、可行组合与先建哪一百页」(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