Cross-Domain Opinion Word Identification with Query-By-Committee Active Learning - Technologies and Applications of Artificial Intelligence - page 339

0.95 0.95

0.85 0.85

Upper bound
Proposed
approach
Random

Upper bound
Proposed
approach
Random

0.75 0.75

ġ
ġġ
ġġġ

ġ
ġġ
ġġġ

0.65 0.65

Lower bound Lower bound

0.55 0.55

0.45 0.45

0.35 0.35

0 100 200 300 400 500 600 700 800 900 1000 0 100 200 300 400 500 600 700 800 900 1000

(a) Movie → Restaurant (b) Hotel → Restaurant

0.9 0.9

0.85 0.85

Upper bound
Proposed
approach
Random
Lower bound

Upper bound
Proposed
approach
Random
Lower bound

0.8 0.8

ġ
ġġ
ġġġ

ġ
ġġ
ġġġ

0.75 0.75

0.7 0.7

0.65 0.65

0.6 0.6

0 100 200 300 400 500 600 700 800 900 1000 0 100 200 300 400 500 600 700 800 900 1000

(c) Restaurant → Movie (d) Hotel → Movie

0.9 0.9

0.85 0.85

0.8 0.8

Upper bound
Proposed
approach
Random

Upper bound
Proposed
approach
Random

0.75 0.75

ġ
ġġ
ġġġ

ġ
ġġ
ġġġ

0.7 0.7

0.65 Lower bound 0.65 Lower bound

0.6 0.6

0.55 0.55

0.5 0.5

0 100 200 300 400 500 600 700 800 900 1000 0 100 200 300 400 500 600 700 800 900 1000

(e) Restaurant → Hotel (f) Movie → Hotel

Fig. 2. Cross-domain performance in all the six source-target domain pairs

Tabl e 3. The ratio of the tokens of the target-domain dataset ( D T ) appearing in the
source-domain dataset ( D S )

Source
Restaurant Movie Hotel

Restaurant - 27.1% 35.8%

Target Movie 32.6% - 33.0%

Hotel 33.5% 25.7% -

our QBC-based method. The stop criteria threshold t is set to 0.003. A denotes
the average number of human annotated sentences in thirty datasets. Table 4
shows the summarized results.