Table 2 Prediction success of the classifiers for gills (A) and digestive gland (B) from the Blue mussel for pairwise comparison of controls (C) with either of the two different exposures to i) 0.5 mg/L crude oil (oil) and ii) 0.5 mg/L crude oil spiked with a mixture of APs and PAHs (0.1 mg/L; sO) for 21 days as validated with an independent sample set not used in model construction. Sample size for learning and testing set, amount of samples attributed to each class, individual and overall classification success in % of testing set is presented.

Class

Learning

Testing

%

C

oil

sO

Tot %

C

34

17

65

11

6

72

oil

44

22

77

5

17

C

34

17

82

14

3

66

sO

37

18

50

9

9

Class

Learning

Testing

%

C

oil

sO

Tot %

C

48

26

85

22

4

76

oil

47

24

67

7

17

C

48

26

77

20

6

80

sO

45

24

83

4

20