Skip to main content

Table 2 Examples of how performance evaluation metrics were calculated

From: Microbial phenomics information extractor (MicroPIE): a natural language processing tool for the automated acquisition of prokaryotic phenotypic characters from text sources

Example #

Character

GSM value

# GSM values

Extracted value

# extracted values

Rigid hit score

Relaxed hit score

1

%G + C

55.2 │ mol%

1

55.2 │ mol%

1

1

1

2

Organic Compounds NOT Used or NOT Hydrolyzed

esculin

1

Neither lactate nor pyruvate

1

0

0

3

Cell Shape

short plump │ rods

1

plump │ rods # short

2

0.5

1

4

Motility

not │ motile by gliding

1

not │ motile

1

0.5

0.5

5

Fermentation Substrates Used

arbutin # salicin # D-raffinose # D-mannose # sucrose # melibiose

6

melibiose # sucrose # D-mannose # D-raffinose # salicin # Most strains ferment arbutin

6

5.5

6

Total

  

10

 

11

7.5

8.5

  1. Rigid and relaxed hit scores measuring the match between extracted values and gold standard matrix (GSM) values, illustrated with examples