HitIct中文无损压缩算法性能评估测试集.pdf

30 3 Vol.30 No.3 2009 3 Journal on Communications March 2009 HitIct 1 2 1,2 2 1. 150001 2. 100190 ANSI HitIct ASCII HitIct Corpus ASCII TP309.3 A 1000-436X(2009)03-0042-06 HitIct: Chinese corpus for the evaluation of lossless compression algorithms 1 2 1,2 2 CHANG Wei-ling , YUN Xiao-chun , FANG Bin-xing , WANG Shu-peng (1. Research Center of Computer Network and Information Security Technology, Harbin Institute of Technology, Harbin 150001, China; 2. Research Center for Information Intelligence and Information Security, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China) Abstract: HitIct, a Chinese corpus for the evaluation of lossless compression algorithms based on ANSI code, was pro- posed. In accordance with the principle of application representativeness, Complementary principle and openness princi- ple, a large number of candidate files were obtained from the Internet, and then average compression ratio, average cor- relation coefficient, compression ratio correlation coefficient and standard deviation were used to select the files that give the most accurate indication of the overall performance of compression algorithms. Experimental results show that this collection has a good representativeness and s