Linear Regression from scratch using Linear Algebra concept

LeastSquares.py

def LinearRegression(x,y,round_digits):
    ''' 
        x: feature vector (Preferably Numpy Array)
        y: dependent vector (Preferably Numpy Array)
        round_digits: number of digits you want to round the final equation to
        This function returns the Ordinary Least Squares Regression model.
    '''
    c = np.ones((len(x[0]),))
    A = []
    A.append(c)
    for col_vec in x:
        A.append(col_vec)
    A = np.asmatrix(A)
    # print(A[0])

    b = np.asarray(y)
    #messed up some dimensions
    A_t = A
    A = A.transpose()
    # print(A.shape)
    ab = np.matmul(A_t,b)
    aa = (np.matmul(A_t,A))
    inverse_aa = np.linalg.inv(aa)
    hat = np.matmul(inverse_aa,ab.transpose())
    hat = np.asarray(hat)
    final_result = f"y = {round(hat[0][0],round_digits)}"
    index = 1
    for i in hat[1:]:
        final_result += f" +{round(i[0],round_digits)}x{index}"
        index+=1
        
    return final_result