Journal of Big Data

Table 2 Error rates for different number of nodes

From: Meta-MapReduce for scalable data mining

Data	Number of computing nodes
	1	5	10	15	20
yeast	0.3071	0.3150	0.3520	0.3565	0.3497
wineRed	0.2351	0.2270	0.2313	0.2332	0.2457
wineWhite	0.2276	0.2274	0.2213	0.2274	0.2274
pendigits	0.0737	0.0693	0.0644	0.0648	0.0625
spambase	0.0680	0.0575	0.0554	0.0575	0.0547
musk	0.0180	0.0270	0.0265	0.0359	0.0380
telescope	0.1608	0.1565	0.1534	0.1507	0.1487
kdd	0.0342	0.0360	0.0300	0.0279	0.0290
isolet	0.1530	0.1367	0.1394	0.1450	0.1535
org	0.1578	0.1457	0.1432	0.1636	0.1690
census	0.0496	0.0492	0.0491	0.0490	0.0482

Lower error rates than those obtained with one node are marked in bold

Back to article page