关于「序贯贝叶斯甄别:似然比账本与沃尔德检验的继续—止损边界」,《真心的信号与时冷时热的陷阱》给出的定量结论是什么?
结论:第二章把每次互动记入似然比账本,示意参数下一次失约的分量相当于 2.52 次兑现,沃尔德序贯检验平均比固定观察期节省 0.459 的互动。
数据口径:书中参数均为示意值;全部计算由书中程序生成;本书不是心理咨询,不宣称任何效果,也不教授任何操控手段。 第二章把每次互动记入似然比账本,示意参数下一次失约的分量相当于 2.52 次兑现,沃尔德序贯检验平均比固定观察期节省 0.459 的互动。
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: 2.52
| step to ledger: kept -0.4353, missed +1.0986 | |||
| one miss outweighs 2.52 kept plans | |||
| t | obs | ledger | P(opportunist) |
| 0 | - | -0.8473 | 0.3000 |
| 1 | 1 | -1.2826 | 0.2171 |
| 2 | 1 | -1.7179 | 0.1521 |
| 3 | 1 | -2.1533 | 0.1040 |
| 4 | 1 | -2.5886 | 0.0699 |
| 5 | 1 | -3.0239 | 0.0464 |
| 6 | 0 | -1.9253 | 0.1273 |
| 7 | 1 | -2.3606 | 0.0862 |
| 8 | 0 | -1.2620 | 0.2206 |
| 9 | 1 | -1.6973 | 0.1548 |
| 10 | 0 | -0.5987 | 0.3546 |
| 11 | 0 | +0.4999 | 0.6224 |
| 12 | 1 | +0.0646 | 0.5161 |
查看完整验算程序
# Chapter 2. p21: a likelihood-ratio ledger over twelve interactions.
# Each interaction is scored 1 if a small stated plan was kept, else 0.
# Illustrative rates: P(kept | genuine) = 0.85, P(kept | opportunist)
# = 0.55 (so a miss has probability 0.15 versus 0.45); prior
# P(opportunist) = 0.3. The ledger adds ln of the ratio
# P(obs | O)/P(obs | G); the posterior follows from Bayes' rule. The
# sequence is a fixed illustrative record, not data.
import math
PG, PO, PRIOR = 0.85, 0.55, 0.3
SEQ = [1, 1, 1, 1, 1, 0, 1, 0, 1, 0, 0, 1]
up_keep = math.log(PO / PG)
up_miss = math.log((1 - PO) / (1 - PG))
print("step to ledger: kept {:+.4f}, missed {:+.4f}".format(up_keep, up_miss))
print("one miss outweighs {:.2f} kept plans".format(-up_miss / up_keep))
print(" t obs ledger P(opportunist)")
llr = math.log(PRIOR / (1 - PRIOR))
print("{:2d} - {:+7.4f} {:.4f}".format(0, llr, PRIOR))
steps = []
for t, x in enumerate(SEQ, 1):
steps.append(up_keep if x else up_miss)
llr = math.log(PRIOR / (1 - PRIOR)) + math.fsum(steps)
post = 1.0 / (1.0 + math.exp(-llr))
print("{:2d} {} {:+7.4f} {:.4f}".format(t, x, llr, post))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《亲密关系早期高代价信号甄别、间歇性强化多巴胺成瘾陷阱与对等底线停时博弈》第 2 章「序贯贝叶斯甄别:似然比账本与沃尔德检验的继续—止损边界」(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-181 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026181,
author = {Fan, Yafu},
title = {Costly-Signal Screening in Early Intimacy, Intermittent-Reinforcement Dopamine Traps and Reciprocal-Floor Stopping Games},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}