SENSEMAP(5WN) manual page

NAME

wn15to16.map, wn15to16mo.map - mapping from noun senses in WordNet 1.5 to corresponding 1.6 senses

WordNet users who have semantically tagged text to senses in version 1.5, or who have statically assigned or used 1.5 senses in other applications should upgrade to WordNet 1.6 if possible. To help users automatically convert 1.5 noun senses to their corresponding 1.6 senses, we provide noun sense mapping information with version 1.6.

The sense mapping was done as follows:

Nouns unique to either database were ignored.
·
Nouns that are monosemous in both databases were found and their sense_key s and synset_offset s were mapped. These sense mappings are in the file 15to16mo.map .
All senses of polysemous nouns in version 1.5 were mapped to senses in version 1.6. Various heuristics were used to evaluate the similarity of 1.5 and 1.6 senses, and a score was assigned to each comparison. For each word, each 1.5 sense was compared to all of the 1.6 senses for the same word, and the 1.6 sense (or senses) with the highest score was deemed the best mapping. These sense mappings are in the file 15to16.map .

File Format

A sense mapping is generally represented by two sense_key synset_offset pairs, one for the 1.5 sense and one for its corresponding 1.6 sense. In 15to16.map , the sense_number is also in each pair (but is not needed in 15to16mo.map , since all monosemous words are assigned sense_number 1). See senseidx(5WN) for a detailed description of these fields.

In both of the following files, a space is the field delimiter unless otherwise noted, and each line is terminated with a newline character.

15to16mo.map

This file contains the mapping of sense keys for nouns that are monosemous in both WordNet 1.5 and 1.6. Although the actual words and sense numbers are the same in both databases, not all sense_keys are the same, and the synset_offset s are different. This file is an alphabetized list of one mapping per line. Each line is of the form:

1.5sense_key;1.5synset_offset 1.6sense_key;1.6synset_offset

15to16.map

This file contains the mapping of sense keys for nouns that are polysemous in WordNet 1.5 and are also found in 1.6. This file is sorted by score from highest score (100) to lowest (0), and then alphabetically within each score. Each line lists all 1.6 sense(s) that the corresponding 1.5 sense maps to with that score. Each line is of the form:

score 1.5sense_info 1.6sense_info [1.6sense_info...]

where sense_info consists of the following three fields:

sense_key;synset_offset;sense_number

SCORES AND STATISTICS

Scores range from 0 to 100, and are an indication of how confident the mapping heuristics are that the senses are the same - a higher score indicates greater reliability in the mapping. The vast majority of senses mapped with a score of 90 or 100.

A total of 32,165 senses for polysemous nouns are mapped in the file 15to16.map , with the following scores:

Score	Count
100	26508
90	4259
80	472
70	316
60	117
50	35
40	75
30	67
20	244
0	72

74,080 monosemous nouns are mapped in the file 15to16mo.map .

There are 107,424 noun senses in WordNet 1.5. A total of 106,245 senses have been mapped to senses in version 1.6. The remaining 1,179 senses represent one or more senses of nouns unique to version 1.5.

NOTES

For several reasons, only noun senses have been mapped. We believe that the vast majority of semantic tagging to WordNet 1.5 senses has involved only nouns, so this should help many users upgrade to version 1.6.

It may seem as if many senses have been deleted from version 1.6, however most of the nouns unique to version 1.5 are collocations or hyphenated strings which are still represented in version 1.6, but with a different combination of hyphens and spaces. Others are spelling or typographical errors that have been corrected.

The number of senses of a polysemous noun in version 1.5 often differs from the number of senses for the same noun in version 1.6. While there will always be a mapping for each 1.5 sense to one or more 1.6 senses, there may be 1.6 senses to which no 1.5 sense is mapped.

WordNet 1.5 nouns not found in either 15to16mo.map or 15to16.map are unique to version 1.5, and therefore cannot be mapped to version 1.6.

ENVIRONMENT VARIABLES

WNHOME: Base directory for WordNet. Unix default is /usr/local/wordnet1.6 , PC default is C:\wn16 , Macintosh default is : .

FILES

All files are in WNHOME/sensemap on Unix platforms, WNHOME\sensemap on PC platforms, and WNHOME:Sensemap on Macintosh platforms.

15to16.map: mapping of polysemous 1.5 noun senses to 1.6 senses
15to16mo.map: mapping of monosemous 1.5 noun senses to 1.6 senses