关于「实证闭环:计分卡、合规模板与预注册方案」,《发酵竞争与定植阈值》给出的定量结论是什么?

结论:第八章给出实证闭环:零身体效果字样的选品说明模板、四周饮食多样性计分卡,以及以 13.1865 为靶的预注册方案。全书含 18 段可运行程序。

数据口径:在书中的示意参数下这个数是 -0.2900 每天,于是胶囊菌株只能在持续服用期间被动携带,携带量 0.0690,停服后半衰期 2.39 天。 书中参数为示意;全部计算由书中程序生成;本书不是医疗建议,不宣称任何效果,自我试验须先咨询执业医师。

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: 13.1865

predicted gain at 6.30 servings for ten weeks13.1865
spread of change effect (SD) people needed
10.01.3187 4.5
15.00.8791 10.2
20.00.6593 18.1
25.00.5275 28.2
30.00.4396 40.6
people smallest gain visible at spread 20
2012.5291
3010.2299
408.8594
607.2337
refutation rule: with 40 people and spread 20, an
upper confidence limit below 13.1865 taxa leaves the
model without the gain it predicts
half intake predicts 5.9171 taxa, so a flat result at
3.00 servings and a gain at 6.30 is the dose-response
pattern the model stands or falls by
查看完整验算程序
# Chapter 8. p82: Sizing the test before running it. The model of
# p42 predicts a richness gain after ten weeks; with an assumed
# spread of individual change, this gives the number of people a
# within-person test needs, the smallest gain a given number can
# see, and the reading that would refute the model.
# Illustrative spreads; the effect comes from the model, not data.
import math

ALPHA, POWER = 0.05, 0.80
POOL, LOSS = 400.0, 0.010
A0, B = 0.0042115, 0.0001862
R0 = 120.0


def phi(z):
    return 0.5 * (1.0 + math.erf(z / math.sqrt(2.0)))


def zq(p, lo=-10.0, hi=10.0):
    for _ in range(200):
        mid = 0.5 * (lo + hi)
        if phi(mid) < p:
            lo = mid
        else:
            hi = mid
    return 0.5 * (lo + hi)


def richness_after(serv, days=70.0):
    a = A0 + B * serv
    eq = a * POOL / (a + LOSS)
    return eq + (R0 - eq) * math.exp(-(a + LOSS) * days)


if __name__ == "__main__":
    z = zq(1.0 - ALPHA / 2.0) + zq(POWER)
    eff = richness_after(6.3) - R0
    print("predicted gain at 6.30 servings for ten weeks: {:.4f}"
          .format(eff))
    print("   spread of change   effect (SD)   people needed")
    for sd in (10.0, 15.0, 20.0, 25.0, 30.0):
        d = eff / sd
        print("   {:16.1f}   {:11.4f}   {:13.1f}"
              .format(sd, d, (z / d) ** 2))
    print("   people   smallest gain visible at spread 20")
    for n in (20, 30, 40, 60):
        print("   {:6d}   {:11.4f}".format(n, z * 20.0 / math.sqrt(n)))
    print("   refutation rule: with 40 people and spread 20, an")
    print("   upper confidence limit below {:.4f} taxa leaves the"
          .format(eff))
    print("   model without the gain it predicts")
    half = richness_after(3.0) - R0
    print("   half intake predicts {:.4f} taxa, so a flat result at"
          .format(half))
    print("   3.00 servings and a gain at 6.30 is the dose-response")
    print("   pattern the model stands or falls by")

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《发酵食品菌群洛特卡—沃尔泰拉竞争、益生菌胶囊定植失败与膳食多样性炎症指标母典》第 8 章「实证闭环:计分卡、合规模板与预注册方案」(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-185 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026185,
  author    = {Fan, Yafu},
  title     = {Lotka-Volterra Competition in Fermented-Food Microbiota, Probiotic-Capsule Colonisation Failure and Dietary-Diversity Inflammation Markers},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论