Skip to main content

Advertisement

Table 2 The average prediction time of CRF++ on single node vs Spark version

From: Recognition of bacteria named entity using conditional random fields in Spark

Data sets (The number of abstracts) (s) Spark version (different numbers of processor cores) (s)
12 24 36 48
2000 362.411 118.479 83.758 75.223 72.375
10,000 1716.569 533.486 325.471 286.723 268.614
20,000 3081.027 964.063 612.743 525.29 517.477
30,000 5207.298 1406.216 883.148 793.282 734.974
40,000 6141.149 1858.607 1168.061 1020.059 966.032
50,000 7956.735 2154.872 1465.193 1243.926 1191.362