Information Theory Coding信息论与编码（英文版） 梁建武 第3章 离散信源新.pptVIP

3.1 Mathematical model and classification of the source Discrete source: outputs the messages which is limited or computable, each time only outputs one message, namely dual messages is not accommodating. Mathematical model ： continuous source： outputs the messages which is limited or computable, each time only outputs one message Mathematical model： Discrete steady source: In the output random sequence , value of each random variable is separate, and the each dimension probability distribution of random vector X does not change along with the time translation. Continuous steady source: In the output random sequence , each value of the random variable is continual, and the each dimension probability density function of random vector X does not change along with the time translation. The discrete source without memory : The discrete source sends out the symbols which do not have the dependence, they are statistical independent. The N expansion source of the discrete source without memory X : the source which is constituted from the N long random sequence output by the discrete source without memory. Mathematical model： N step space of source space X source with memory: to the output random sequence X, each random variable has the dependent relations, but the memory length is limited. m step Markov source: The symbol which source each time sends out is only related with the previous m symbols, has nothing to do with the front mark. periodic Markov source: The above conditional probability has nothing to do with the time beginning i Stochastic profile source: The source output messages have the continual function both in the time (or space) and in the value. 3.2 The discrete source without memory The discrete source only concern with one random event, it can be expressed by discrete random variable. Mathematical model: X，Y，Z are on behalf of the random v