Table 4.
Strategy |
Abbr →Exp |
Exp →Abbr |
Syn |
||||||
---|---|---|---|---|---|---|---|---|---|
RI | RP | Result | RI | RP | Result | RI | RP | Result | |
RI ⊂ RP30 |
RI_8 |
RP_8_sw |
0.38 |
RI_8 |
RP_8 |
0.30 |
RI_8 |
RP_8 |
0.39 |
|
|
|
|
RI_4 |
RP_4_sw |
0.30 |
RI_8 |
RP_8 |
0.38 |
RP ⊂ RI30 |
RI_20 |
RP_4_sw |
0.35 |
RI_20 |
RP_4_sw |
|
RI_8 |
RP_8_sw |
|
|
|
|
|
|
|
|
RI_20 |
RP_2_sw |
|
RI + RP | RI_4 | RP_4_sw | 0.42 | RI_4 |
RP_4_sw |
0.32 | RI_20 | RP_4_sw | 0.40 |
RI_8 | RP_8_sw |
Results (recall, top ten) of the best configurations for each model and model combination on the three tasks. The configurations are described according to the following pattern: model_windowSize. For RP, sw means that stop words are retained in the model.