TY - JOUR
T1 - Combining Forward Compression with PPM
AU - Avrunin, Rachel Mustakis
AU - Klein, Shmuel T.
AU - Shapira, Dana
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.
PY - 2022/5
Y1 - 2022/5
N2 - A new Forward Looking variant of dynamic Huffman or arithmetic encoding has been recently proposed, that provably always performs better than the corresponding general static encoding schemes, as far as the net compressed file, without the necessary header, is concerned. The current paper suggests to integrate the Forward Looking paradigm with the well-known adaptive PPM—Prediction by Partial Matching algorithm. This combination, that attempts to predict the following character based on the context that has already occurred in past, but uses its knowledge of the exact frequencies in the future, is empirically shown to enhance the prediction capability, and therefore to improve the compression efficiency.
AB - A new Forward Looking variant of dynamic Huffman or arithmetic encoding has been recently proposed, that provably always performs better than the corresponding general static encoding schemes, as far as the net compressed file, without the necessary header, is concerned. The current paper suggests to integrate the Forward Looking paradigm with the well-known adaptive PPM—Prediction by Partial Matching algorithm. This combination, that attempts to predict the following character based on the context that has already occurred in past, but uses its knowledge of the exact frequencies in the future, is empirically shown to enhance the prediction capability, and therefore to improve the compression efficiency.
KW - Arithmetic coding
KW - Lossless compression
KW - PPM
UR - http://www.scopus.com/inward/record.url?scp=85128722367&partnerID=8YFLogxK
U2 - 10.1007/s42979-022-01121-0
DO - 10.1007/s42979-022-01121-0
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:85128722367
SN - 2662-995X
VL - 3
JO - SN Computer Science
JF - SN Computer Science
IS - 3
M1 - 239
ER -