Then the paper also builds multi-order context adaptive probability model, which enables the shortened coded character almost approach its entropy rate.

• 建立了基于上下文的多阶自适应的概率模型,使其符号的压缩码长尽量的接近其熵值。