摘要: This paper discusses a switching method which can be used to combine two sequential universal source coding algorithms. The treats these algorithms as black-boxes and only use their estimates of the probability distributions for consecutive symbols sequence. Three weighting based on this are presented. Empirical results show that all three give performance better than they combine.
sci-hub.se 参考文章(4)
S. Bunton, A percolating state selector for suffix-tree context models data compression conference. pp. 32- 41 ,(1997) , 10.1109/DCC.1997.581957
G. Langdon, A note on the Ziv - Lempel model for compressing individual sequences (Corresp.) IEEE Transactions on Information Theory. ,vol. 29, pp. 284- 287 ,(1983) , 10.1109/TIT.1983.1056645
F.M.J. Willems, Coding for a binary independent piecewise-identically-distributed source IEEE Transactions on Information Theory. ,vol. 42, pp. 2210- 2217 ,(1996) , 10.1109/18.556608
F.M.J. Willems, Y.M. Shtarkov, T.J. Tjalkens, The context-tree weighting method: basic properties IEEE Transactions on Information Theory. ,vol. 41, pp. 653- 664 ,(1995) , 10.1109/18.382012