Skip to main content

Table 4 Comparisons of SCSM and BM25, with MAP, RP and P@5, P@20, NDCG@5, and NDCG@20 on WT2G collection

From: Salient context-based semantic matching for information retrieval

Corpus

Methods

MAP

RP

P@5

P@20

NDCG@5

NDCG@20

WT2G

BM25

0.313

0.340

0.532

0.391

0.542

0.470

 

CSSMC

0.368

0.378

0.600

0.428

0.616

0.521

  

+ 17.82%

+ 11.09%

+ 12.78%

+ 9.46%

+ 13.63%

+ 10.88%

 

SCSMlf

0.370

0.383

0.592

0.435

0.611

0.526

  

+ 18.27%

+ 12.77%

+ 11.28%

+ 11.25%

+ 12.63%

+ 11.97%

 

SCSMgf

0.370

0.381

0.616

0.426

0.628

0.521

  

+ 18.01%

+ 12.03%

+ 15.79%

+ 8.95%

+ 15.88%

+ 11.05%

  1. Boldface numbers indicate statistically significant improvements over BM25 by permutation test