Skip to main content

Table 5 Best scoring patterns found in both experiments

From: Expressive modeling for trusted big data analytics: techniques and applications in sentiment analysis

Pattern

G

O

#C

Experiment 1

   

[(h u i l e n,v e r b p r e s s g)]

–

–

7

[(t e,p a r t t e),(, a d j)]

–

–

3

[(j a m m e r,v e r b p r e s s g)]

–

–

2

[(j a m m e r,v e r b p r e s s g),(∗,∗),

+

+

1

(a f g e l o p e n,v e r b p a p a)]

   

Experiment 2

   

[(g o e d e,a d j),(t i j d e n,n o u n p l)]

=

+

2

[(s l e c h t e,a d j),(t i j d e n,n o u n p l)]

=

+

2

[(s t o t t e r d,n o u n s g)]

–

+

1

  1. G - pattern group (positive +, negative – and neutral =), O - operation (add +, remove -) and #C - number of corrections