You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Computing intervals. There are a number of options for computing intervals which can impact efficiency, details are in the CIKM paper. The options are:
Eager plane sweep
Lazy plane sweep
Traversal of inverted file (with positional information)
Traversal of direct file
Reported in the CIKM paper is that traversal over the inverted file works best for shorter documents, while using the direct file serves better with longer documents.
An ideal solution would be to support all combinations of methods. However currently there is no support for positional inverted index files, so that would need to be added for inverted file traversal.
With this in mind a first pass on implementing this will use the direct file approach and consider the inverted file approach afterwards.
The text was updated successfully, but these errors were encountered:
From Efficient and Effective Higher Order Proximity Modeling. The intervals for L2p, Lkp and Lkfp methods are computed using lazy/eager plane sweep Lu et al, (CIKM 15).
Computing intervals. There are a number of options for computing intervals which can impact efficiency, details are in the CIKM paper. The options are:
Reported in the CIKM paper is that traversal over the inverted file works best for shorter documents, while using the direct file serves better with longer documents.
An ideal solution would be to support all combinations of methods. However currently there is no support for positional inverted index files, so that would need to be added for inverted file traversal.
With this in mind a first pass on implementing this will use the direct file approach and consider the inverted file approach afterwards.
The text was updated successfully, but these errors were encountered: