关于「不必全盘抽样:篮子报价、稀疏暗折与一范数的精确重构」,《稀疏探针与暗折重构》给出的定量结论是什么?

结论:第一章核对并修正立项命题的出处,用分数精确的单纯形法以 12 次篮子报价还原 40 件商品里的 3 笔暗折,而最小二乘把它们摊满全表,最坏一格偏离 26.73 元;并量出排程常数,报价占商品数的比例可降到 0.0825。

数据口径:只要暗折是稀疏的,把商品随机组合成购物车、只读结算总额,所需报价次数大致与暗折数乘以商品数对暗折数之比的对数成正比,远少于商品数;在非负约束下极小化一范数,无噪声时得到的是精确的表。 书中商品、金额、成本与竞品均为示意;全部计算由书中程序生成;本书是非个人化经营参考,不是法律意见。

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: 12, 26.73, 40

cells n40, hidden discounts k = 3, basket quotes m = 12
full census40 single-item quotes
basket sizes25 15 21 20 13 19 19 22 15 24 14 15
discount totals y30 0 30 18 0 35 12 47 35 65 35 12
cell true L1 (exact) min-L2
718 18 5.49
2135 35 8.27
3312 12 3.12
L1 equals truth in all 40 cells, as fractions: True
L1 nonzero cells3
min-L2 nonzero cells40, cells off by 1 yuan or more: 31
min-L2 worst cell error26.73 yuan
simplex pivots (exact, Bland's rule)50
first q quotes only, exact? 3:n 4:n 5:n 6:n 7:n 8:n 9:n 10:Y 11:Y 12:Y

核心公式

\[y = \sum_{j \in \text{basket}} p_j - T = \sum_{j \in \text{basket}} x_j .\]
查看完整验算程序
# Chapter 1. p11: 12 basket quotes recover 3 hidden discounts among 40.
# Model: a competitor lists n = 40 items at public shelf prices; under
# one public checkout condition, k = 3 items carry a hidden discount
# (whole yuan). A probe is a basket quote that shows only the total;
# shelf total minus quoted total = sum of hidden discounts in the
# basket, so y = A x with A a 0/1 basket matrix. We compare basis
# pursuit (min sum x, A x = y, x >= 0; exact simplex in fractions)
# with the minimum-L2 answer x = A'(AA')^-1 y. Illustrative numbers;
# baskets are drawn with a fixed seed, each item in with chance 1/2.
import math
import random
from fractions import Fraction as Fr

def esum(v):
    s = Fr(0)
    for t in v:
        s += t
    return s


def lp_min(A, b, c, exact=False):
    # min c.x subject to A x = b, x >= 0: two-phase tableau simplex.
    # b is nudged to b + A d (d_j > 0, tiny, distinct) so that no
    # pivot is degenerate; the final basis is re-applied to the true
    # b through B^-1. Exact mode uses fractions and Bland's rule.
    m, n = len(A), len(A[0])
    cv = Fr if exact else float
    add = esum if exact else math.fsum
    tol = 0 if exact else 1e-9
    big = 1 + max(abs(v) for v in b)
    d = [Fr(1, 2 ** (30 + j)) if exact else
         1e-7 * big * (0.5 + (0.6180339887 * (j + 1)) % 1.0)
         for j in range(n)]
    T, sg = [], []
    for i in range(m):
        s = -1 if b[i] < 0 else 1
        row = [cv(s * v) for v in A[i]] + [cv(0)] * (m + 1)
        row[n + i] = cv(1)
        row[-1] = cv(abs(b[i])) + s * add([A[i][j] * d[j]
                                           for j in range(n)])
        T.append(row)
        sg.append(s)
    basis = [n + i for i in range(m)]
    W = n + m

    def pivot(r, q, z):
        pv = T[r][q]
        T[r] = [v / pv for v in T[r]]
        pr = T[r]
        nz = [j for j in range(W + 1) if pr[j] != 0]
        for row in T + [z]:
            f = row[q]
            if row is not pr and f != 0:
                for j in nz:
                    row[j] -= f * pr[j]
        basis[r] = q

    def run(z, stop):
        piv = 0
        while not (stop and -z[W] <= tol):
            q, best = -1, -1e-7
            for j in range(n):
                if exact and z[j] < 0:
                    q = j
                    break
                if not exact and z[j] < best:
                    q, best = j, z[j]
            if q < 0:
                return piv
            r, rt = -1, None
            for i in range(m):
                if T[i][q] > tol:
                    t = T[i][W] / T[i][q]
                    if rt is None or t < rt:
                        r, rt = i, t
            pivot(r, q, z)
            piv += 1
        return piv

    z = [cv(0)] * (W + 1)
    for j in list(range(n)) + [W]:
        z[j] = -add([T[i][j] for i in range(m)])
    piv = run(z, True)
    if -z[W] > (0 if exact else 1e-6 * big):
        return None, None, piv
    z = [cv(v) for v in c] + [cv(0)] * (m + 1)
    for i in range(m):
        if basis[i] < n and z[basis[i]] != 0:
            f = z[basis[i]]
            z = [z[j] - f * T[i][j] for j in range(W + 1)]
    piv += run(z, False)
    x = [cv(0)] * n
    for i in range(m):
        if basis[i] < n:
            x[basis[i]] = add([T[i][n + j] * abs(b[j])
                               for j in range(m)])
    lam = [-sg[j] * z[n + j] for j in range(m)]
    return x, lam, piv


def solve_exact(G, h):
    k = len(G)
    M = [list(G[i]) + [h[i]] for i in range(k)]
    for c in range(k):
        p = next(i for i in range(c, k) if M[i][c] != 0)
        M[c], M[p] = M[p], M[c]
        for i in range(k):
            if i != c and M[i][c] != 0:
                f = M[i][c] / M[c][c]
                M[i] = [M[i][j] - f * M[c][j] for j in range(k + 1)]
    return [M[i][k] / M[i][i] for i in range(k)]


n, m = 40, 12
truth = {7: 18, 21: 35, 33: 12}
x = [truth.get(j, 0) for j in range(n)]
rng = random.Random(20261113)
A = [[1 if rng.random() < 0.5 else 0 for _ in range(n)]
     for _ in range(m)]
y = [int(math.fsum(A[i][j] * x[j] for j in range(n))) for i in range(m)]
xb, lam, piv = lp_min(A, y, [1] * n, exact=True)
G = [[Fr(esum([Fr(A[p][j] * A[q][j]) for j in range(n)]))
      for q in range(m)] for p in range(m)]
mu = solve_exact(G, [Fr(v) for v in y])
x2 = [esum([A[i][j] * mu[i] for i in range(m)]) for j in range(n)]
err2 = [abs(float(x2[j]) - x[j]) for j in range(n)]
print("cells n = %d, hidden discounts k = %d, basket quotes m = %d"
      % (n, len(truth), m))
print("full census: %d single-item quotes" % n)
print("basket sizes: %s" % " ".join(str(int(math.fsum(r))) for r in A))
print("discount totals y: %s" % " ".join(str(v) for v in y))
print("cell  true   L1 (exact)   min-L2")
for j in sorted(truth):
    print("%4d  %4d   %10s   %6.2f" % (j, x[j], xb[j], float(x2[j])))
print("L1 equals truth in all %d cells, as fractions: %s"
      % (n, xb == [Fr(v) for v in x]))
print("L1 nonzero cells: %d" % len([v for v in xb if v != 0]))
print("min-L2 nonzero cells: %d, cells off by 1 yuan or more: %d"
      % (len([v for v in x2 if v != 0]),
         len([e for e in err2 if e >= 1])))
print("min-L2 worst cell error: %.2f yuan" % max(err2))
print("simplex pivots (exact, Bland's rule): %d" % piv)
row = []
for mm in range(3, m + 1):
    xs, _, _ = lp_min(A[:mm], y[:mm], [1] * n, exact=True)
    row.append("%d:%s" % (mm, "Y" if xs == [Fr(v) for v in x] else "n"))
print("first q quotes only, exact? " + " ".join(row))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《压缩感知一范数极小化、神秘访客稀疏探针与竞品暗折矩阵精确重构母典》第 1 章「不必全盘抽样:篮子报价、稀疏暗折与一范数的精确重构」(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-242 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026242,
  author    = {Fan, Yafu},
  title     = {Compressed Sensing and L1 Minimisation: Sparse Mystery-Shopper Probes and the Exact Recovery of Competitors' Hidden-Discount Matrices},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论