Getting an error “could not broadcast input array from shape (252,4) into shape (4)” in optimization

问题

I a relatively new to python scipy library. I was trying to use the scipy.optimize to find the maximum value of the sharpe() function in the following code

def sharpe(dr, wts):
    portfolio=np.ones(dr.shape[0])
    dr[0:]=wts*dr[0:]
    sharpe_ratio=-np.mean(np.sum(dr, axis=1))/np.std(np.sum(dr, axis=1))
    return sharpe_ratio

def wts_con(wts):
    return wts[0]+wts[1]+wts[2]+wts[3]-1

def sharpe_optimize(dr, sharpe_func):
    wts_guess=np.array([0.25,0.25,0.25,0.25])

    con=[{"type":"eq", "fun":wts_con}]
    bnds=((0,1),(0,1),(0,1),(0,1))

    result=spo.minimize(sharpe_func, wts_guess, args=(dr,), method="SLSQP", constraints=con, bounds=bnds, options={"disp":True})
   return result

In the above code dr is a size(252,4) array and wts is a size(4) array

I am getting the following error in line 3 when I call the sharpe_optimize() function

could not broadcast input array from shape (252,4) into shape (4)

回答1:

I think you need to swap the arguments in your definition of the sharpe function. It is defined as sharpe(dr,wts) but then it looks like you call minimize as sharpe(wts,dr) based on your use of args. Edit: I have just seen that this is pointed out in the above comments!

来源：https://stackoverflow.com/questions/39310546/getting-an-error-could-not-broadcast-input-array-from-shape-252-4-into-shape