R语言利用三个矩阵对应值调用uniroot求解方程根的实现问题

核心思路 #

你需要对每一行的a、b参数，分别对应三个年份的pop、event取值单独求解方程，原写法直接传入整个矩阵会导致uniroot无法识别单一求解目标，所以我们通过逐行逐列遍历调用uniroot实现需求，同时也可以直接用解析解计算（效率更高）。

完整可运行代码 #

# 构造三个矩阵（兼容base R写法，无需额外依赖）
df <- data.frame(
  a = c(-5.8221, -6.4678, -6.1213, -5.4518, -5.3527, -5.3517, -5.3169),
  b = c(.0234, .0199, .0162, .0151, .0148, .0138, .0123)
)
a_b <- as.matrix(df[1:7, 1:2])

df1 <- data.frame(
  "2000" = c(716209.27, 642504.69, 644640.11, 632239.35, 552437.34, 554395.34, 539358.87),
  "2005" = c(653040.76, 637752.95, 651635.40, 647572.55, 641183.99, 570938.83, 554092.75),
  "2010" = c(5643.6, 876123, 665412.5, 98643, 89012, 776321, 8863),
  check.names = FALSE
)
pop <- as.matrix(df1[1:7, 1:3])

df2 <- data.frame(
  "2000" = c(9968.50418, 2605.16992, 2937.01261, 4286.23376, 5718.80904, 4843.19769, 5077.52440),
  "2005" = c(6037.16994, 2336.72681, 2391.55031, 4085.62207, 5000.77218, 4425.02331, 5097.25052),
  "2010" = c(19123.45, 87765.54, 102345.44, 98000, 12396, 90076, 10133),
  check.names = FALSE
)
event <- as.matrix(df2[1:7, 1:3])

# 初始化结果矩阵，和pop同维度、同列名
result <- matrix(NA, nrow = nrow(pop), ncol = ncol(pop), dimnames = dimnames(pop))

# 逐行逐列调用uniroot求解
for (i in seq_len(nrow(a_b))) {
  a <- a_b[i, 1]
  b <- a_b[i, 2]
  for (j in seq_len(ncol(pop))) {
    e_val <- event[i, j]
    p_val <- pop[i, j]
    # 定义单目标求解函数
    f <- function(x) {
      e_val - exp(a + b * x) * p_val
    }
    # 调用uniroot求解，存根到结果矩阵
    result[i, j] <- uniroot(f, c(-10000, 10000), extendInt = "yes")$root
  }
}

# 保留4位小数输出，和你预期格式一致
round(result, 4)

# （可选）解析解直接计算，效率更高，结果完全一致
# result_analytical <- (log(event/pop) - a_b[,1])/a_b[,2]
# round(result_analytical, 4)

输出结果 #

运行代码后会得到你需要的格式：

2000    2005    2010
[1,] 66.1349 48.6493 300.9612
[2,] 48.2373 43.1457 209.3950
[3,] 45.0617 31.7133 262.2990
[4,] 30.3271 25.5657 360.6132
[5,] 52.8459 33.7145 228.4664
[6,] 44.3041 35.6299 231.7237
[7,] 52.9546 51.0787 443.1554

内容的提问来源于stack exchange，提问作者mehmo