We are using the Levenshtein algorithm \u0026lt; http://en.wikipedia.org/wiki/ Levenshtein_distance > as a precursor to the disambiguation of the names of English researchers in biomedicine.
In the initial pilot, which takes the original ISI published from 1990 to 1994 the number of signatures is 44363.
Since we have denied funding for fairly decent servers or computers, the calculation for each year employs two full days, keeping in mind that we move with variable thresholds.
Does anyone know how to speed up the algorithm? Or, put another way: Does anyone know of a variant of the algorithm more efficient?
In the initial pilot, which takes the original ISI published from 1990 to 1994 the number of signatures is 44363.
Since we have denied funding for fairly decent servers or computers, the calculation for each year employs two full days, keeping in mind that we move with variable thresholds.
Does anyone know how to speed up the algorithm? Or, put another way: Does anyone know of a variant of the algorithm more efficient?
0 comments:
Post a Comment