Skip to main content

Table 4 Stacking experiments for the ID track

From: Combining joint models for biomedical event extraction

System

Recall

Precision

F1

UMass

46.2

51.1

48.5

Stanford (1N)

46.9

50.2

48.5

Stanford (1P)

44.4

47.7

46.0

Stanford (2N)

45.0

54.8

49.4

Stanford (2P)

46.6

49.2

47.8

Stanford (1N, reranked)

47.5

51.4

49.4

Stanford (1P, reranked)

47.9

49.2

48.5

Stanford (2N, reranked)

45.7

52.3

48.8

Stanford (2P, reranked)

49.6

49.9

49.8

Stanford (all, reranked)

48.9

51.6

50.2

UMass←Stanford (1N)

45.8

51.6

48.5

UMass←Stanford (1P)

47.6

52.8

50.0

UMass←Stanford (2N)

45.4

52.4

48.6

UMass←Stanford (2P)

49.1

52.6

50.7

UMass←Stanford (all)

47.6

54.3

50.7

UMass←Stanford (2P, Conj)

48.0

53.2

50.4

  1. Results on the development set for the ID track.