使用fmincg进行逻辑回归

Question

我正在尝试使用fmincg optimiser在python中实现逻辑回归，但是我不断遇到多个错误。有更好的方法吗？

<form @submit.prevent="editmode ? updateStudent() : insertStudent()">
                         <div class="modal-body">
                            <!-- Students registration form start from here-->
                                <div class="form-group">
                                <input v-model="form.admission_no" type="text" name="admission_no" placeholder="Admission Number"
                                    class="form-control" :class="{ 'is-invalid': form.errors.has('admission_no') }">
                                <has-error :form="form" field="admission_no"></has-error>
                                </div>

def cost(theta,X_train,y_train,L1): theta0 = theta[0]; theta1 = theta[1:]; z =-(X_train@theta); hyp = np.ones(z.shape)/(np.ones(z.shape)+np.exp(z)); temphyp1 = y_train.transpose()@np.log(hyp); temphyp2 = (1-y_train.transpose())@np.log(1-hyp); J = -(1/m*(temphyp1+temphyp2))+((L1/(2*m))*np.sum(theta1**2)) grad0 = -(1/m)*X_train[:,0].T@(hyp-y_train); grad1 = -(1/m)*X_train[:,1:].T@(hyp-y_train)+((L1/(m))*np.sum(theta1)); grad = np.vstack((grad0,grad1)) return J.flatten(), grad.flatten()