Compression-based classification of biological sequences and structures via the Universal Similarity Metric: experimental assessment

Table 2 CK-36-REL

CK-36-REL	UCD		NCD		CD
	UPGMA	NJ	UPGMA	NJ	UPGMA	NJ
Gzip	0.8870	0.7820	0.8870	0.8676	0.9400	0.8195
Bzip2	0.9030	0.8197	0.9030	0.7970	0.9030	0.8096
PPMd16	0.9030	0.9030	0.9030	0.7820	0.9030	0.9015
PPMd8	0.9030	0.7820	0.9030	0.7820	0.9030	0.7600
PPMd4	0.9030	0.7820	0.9030	0.7820	0.9030	0.8478
PPMd2	0.8881	0.7450	0.9030	0.8329	0.9030	0.7600
Huffman	0.8663	0.7700	0.8663	0.7700	0.8274	0.7658
Ac fast	0.8580	0.7639	0.8580	0.7639	0.9030	0.7777
Rc fast	0.8580	0.7639	0.8400	0.7820	0.8734	0.7288
Ac med.	0.8881	0.7820	0.8881	0.7820	0.8734	0.7777
Rc med.	0.8580	0.7440	0.8580	0.7820	0.8274	0.8212
Ac slow	0.8699	0.7820	0.8699	0.7820	0.8022	0.7658
Rc slow	0.8706	0.8706	0.8706	0.7440	0.7552	0.7658
BwtRleHuff	0.8706	0.7558	0.8706	0.7440	0.9030	0.7694
BwtMtfRleHuff	0.8500	0.8542	0.8518	0.7558	0.8706	0.7275
BwtRleAc fast	0.8881	0.7440	0.8881	0.7820	0.9030	0.7694
BwtMtfRleAc fast	0.8881	0.8542	0.8881	0.8850	0.9030	0.7943
BwtRleRc fast	0.8518	0.8923	0.8518	0.8542	0.9030	0.7694
BwtMtfRleRc fast	0.8706	0.8188	0.8706	0.8146	0.9030	0.7718
BwtRleRc med.	0.8500	0.7440	0.8500	0.7440	0.8706	0.7970
BwtMtfRleRc med.	0.8518	0.7558	0.8706	0.8188	0.9030	0.8863
BwtRleRc slow	0.8881	0.7440	0.8881	0.7398	0.8500	0.7694
BwtMtfRleRc slow	0.8881	0.7558	0.8881	0.8542	0.8500	0.7690
BwtWavelet	0.8392	0.9084	0.8400	0.7898	0.8458	0.7288

Experimental results for the CK-36-REL data set, with the UCD (left), NCD (middle), and CD (right) distance. For each compression algorithm, we report the F-measure for both UPGMA and NJ methods.

ISSN: 1471-2105