随机变量序列基于指数分布的强偏差定理

Strong deviation theorems of random sequence based on exponential distribution

利用样本相对熵作为一般非负连续型随机变量序列相对于服从指数分布的独立但不同分布随机变量序列偏差的一种随机性度量，运用鞅理论及分析方法，研究了一类随机变量序列加权和的用不等式表示的极限定理。

The notion of sample relative entropy, as a random measure of the deviation of a sequence of nonnegative continuous random sequence from independent random sequence with exponential distribution, is introduced. By using the martingale theory and analytic method, some strong limit theorems represented by inequality of weighted random sequence are studied, i.e. strong deviation theorems.