关于「交替最小二乘:把一个难题拆成一串小方程组」,《稀疏购买记录的偏好补全》给出的定量结论是什么?

结论:观测格子共 343 个。岭参数取 0.001 时,训练误差是 0.3886,检验误差却高达 4.2879:几乎不加惩罚,记录少的顾客与商品被拟合到噪声里去,预测在未观测格子上大幅外飘。这一行的训练误差反而比岭参数 0.100 时的 0.3361 还高,说明几乎不加惩罚时交替最小二乘也收敛得不干净。岭参数取 1.000 时检验误差最低,是 0.7649;再往上加,惩罚压过了数据,训练与检验误差一起上升,到 30.000 时两者分别是 1.7089 与 1.7581,模型已经接近什么也没学到。

数据口径:书中顾客数、商品数、口味维度与购买概率均为示意参数;全部计算由书中程序生成;本书属于非个人化经营建议,推荐流程只用店铺自有数据,不外传个人数据,不做跨平台追踪。 第七章考辨《周易·系辞上》「方以类聚,物以群分」与斯皮尔曼 1904 年的一般因素,指出纯分群与混合成员下两种模型各有胜负。

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: 0.001, 0.100, 0.3361, 0.3886, 0.7649, 1.000

observed 343 cells, noise sd 0.5, rank 3
lambdatrain RMSEtest RMSE
0.0010.38864.2879
0.1000.33611.2052
1.0000.39860.7649
3.0000.64281.0035
10.0001.36451.5411
30.0001.70891.7581
查看完整验算程序
# Chapter 2. p22: the ridge penalty with noisy scores.
# Model: 40 customers x 30 products, true rank 3 (Gaussian factors),
# each observed score carries Gaussian noise with sd 0.5; 30 percent
# of cells observed (illustrative). ALS at rank 3 with ridge lambda;
# train RMSE is on observed noisy scores, test RMSE compares the
# predictions on unobserved cells with the noise-free truth.
import math
import random


def dot(a, b):
    return math.fsum(x * y for x, y in zip(a, b))


def solve(A, b):
    # Gaussian elimination with partial pivoting for a small system
    n = len(b)
    M = [list(A[i]) + [b[i]] for i in range(n)]
    for k in range(n):
        p = max(range(k, n), key=lambda i: abs(M[i][k]))
        M[k], M[p] = M[p], M[k]
        for i in range(k + 1, n):
            f = M[i][k] / M[k][k]
            M[i] = [M[i][j] - f * M[k][j] for j in range(n + 1)]
    x = [0.0] * n
    for i in range(n - 1, -1, -1):
        s = math.fsum(M[i][j] * x[j] for j in range(i + 1, n))
        x[i] = (M[i][n] - s) / M[i][i]
    return x


def update(trip, F, r, lam, prior=None):
    # weighted ridge step for one row: minimise
    # sum w (y - f.x)^2 + lam |x - prior|^2 over x
    pr = prior if prior is not None else [0.0] * r
    A = [[math.fsum(w * F[j][a] * F[j][b] for j, _, w in trip)
          + (lam if a == b else 0.0) for b in range(r)]
         for a in range(r)]
    y = [math.fsum(w * F[j][a] * v for j, v, w in trip) + lam * pr[a]
         for a in range(r)]
    return solve(A, y)


def als(nu, ni, obs, r, lam, iters, seed, V=None, trace=None):
    # alternating least squares on (u, i, value, weight) tuples
    rng = random.Random(seed)
    if V is None:
        V = [[rng.gauss(0.0, 1.0) for _ in range(r)] for _ in range(ni)]
    bu = [[] for _ in range(nu)]
    bi = [[] for _ in range(ni)]
    for u, i, x, w in obs:
        bu[u].append((i, x, w))
        bi[i].append((u, x, w))
    U = None
    for t in range(iters):
        U = [update(bu[u], V, r, lam) for u in range(nu)]
        V = [update(bi[i], U, r, lam) for i in range(ni)]
        if trace is not None:
            trace.append((t + 1, U, V))
    return U, V


def factors(n, r, rng):
    return [[rng.gauss(0.0, 1.0) for _ in range(r)] for _ in range(n)]


def product(A, B):
    return [[dot(a, b) for b in B] for a in A]


def mask(nu, ni, p, rng):
    return [(u, i) for u in range(nu) for i in range(ni)
            if rng.random() < p]


def relerr(M, U, V, cells):
    num = math.fsum((dot(U[u], V[i]) - M[u][i]) ** 2 for u, i in cells)
    den = math.fsum(M[u][i] ** 2 for u, i in cells)
    return math.sqrt(num / den)


def rmse(M, U, V, cells):
    num = math.fsum((dot(U[u], V[i]) - M[u][i]) ** 2 for u, i in cells)
    return math.sqrt(num / len(cells))

rng = random.Random(20261113)
nu, ni, r = 40, 30, 3
M = product(factors(nu, r, rng), factors(ni, r, rng))
seen = mask(nu, ni, 0.3, rng)
ss = set(seen)
rest = [(u, i) for u in range(nu) for i in range(ni) if (u, i) not in ss]
noisy = {c: M[c[0]][c[1]] + rng.gauss(0.0, 0.5) for c in seen}
obs = [(u, i, noisy[(u, i)], 1.0) for u, i in seen]
print("observed {} cells, noise sd 0.5, rank 3".format(len(seen)))
print("   lambda   train RMSE   test RMSE")
for lam in (0.001, 0.1, 1.0, 3.0, 10.0, 30.0):
    U, V = als(nu, ni, obs, r, lam, 60, 3)
    tr = math.sqrt(math.fsum((x - dot(U[u], V[i])) ** 2
                             for u, i, x, _ in obs) / len(obs))
    print("  {:7.3f}     {:.4f}      {:.4f}".format(
        lam, tr, rmse(M, U, V, rest)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《低秩矩阵补全与交替最小二乘、稀疏购买记录推断未表态偏好母典》第 2 章「交替最小二乘:把一个难题拆成一串小方程组」(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-328 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026328,
  author    = {Fan, Yafu},
  title     = {Low-Rank Matrix Completion and Alternating Least Squares: Inferring Unstated Preferences from Sparse Purchase Records},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论