Empirical bayes does not converge if few datapoint (<10)

[mathrip]

Need to add an early stopping in it_sol() function

def it_sol(sdat, g_hat, d_hat, g_bar, t2, a, b, conv=0.0001, robust=False):
    n = (1 - np.isnan(sdat)).sum(axis=1)
    g_old = g_hat.copy()
    d_old = d_hat.copy()
    ones = np.ones((1, sdat.shape[1]))

    change = 1
    change_old = 1
    count = 0
    while change > conv:
        g_new = np.array(postmean(g_hat, g_bar, n, d_old, t2))
        
        if robust:
            sum2 = n*biweight_midvar(sdat,center=g_new.reshape((g_new.shape[0], 1)),axis=1)
            # sum2 = n*(1.482602218505602*np.median(abs(sdat - np.dot(g_new.reshape((g_new.shape[0], 1)), ones)), axis = 1)) ** 2
        else:
            sum2 = ((sdat - np.dot(g_new.reshape((g_new.shape[0], 1)), ones)) ** 2).sum(
                axis=1
            )
        
        d_new = postvar(sum2, n, a, b)
        # change = max(
        #     (abs(g_new - g_old.item()) / g_old.item()).max(),
        #     (abs(d_new - d_old) / d_old).max(),
        # )
        change = max(max(abs(g_new - g_old) / g_old), max(abs(d_new - d_old) / d_old))
        # print(max(abs(g_new - g_old) / g_old), "," ,max(abs(d_new - d_old) / d_old))
        
        if count > 30:
            if change > change_old:
                print('[neuroCombat WARNING] Empirical Bayes step failed to converge after 30 iterations, using estimate before change between iterations increases.')
                break

        g_old = g_new  # .copy()
        d_old = d_new  # .copy()
        change_old = change
        count = count + 1
    adjust = (g_new, d_new)
    return adjust