# 定义目标函数 def func(x): return np.sum(x ** 2)

替换后的代码如下：

function dz = Fish_SEIR(t, z, k)     
    beta1 = k(1);     
    mu1 = k(2);     
    gamma1 = k(3);     
    delta1 = k(4);     
    xi = k(5);     
    dz = zeros(5, 1);     
    dz(1) = -beta1 * z(1) * z(3) - xi * z(1);     
    dz(2) = beta1 * z(1) * z(3) + xi * z(1) - mu1 * z(2);     
    dz(3) = mu1 * z(2) - gamma1 * z(3);     
    dz(4) = gamma1 * z(3) - delta1 * z(4);     
    dz(5) = delta1 * z(4); 
end   

function y = Func_to_optimize(k0, z0, tspan0)     
    load('fishdata.mat')     
    T = tspan0(1):0.1:tspan0(2);     
    [t, y] = ode45(@Fish_SEIR, T, z0, [], k0);     
    cdata = cumsum(data);     
    e = y(1:10:end) - cdata;     
    y = sum(e.^2); 
end

这里将原来的函数名 `Diff_report` 改为了 `Func_to_optimize`，并且在函数体内计算了模拟结果与数据之间的残差平方和，作为目标函数的返回值。