Jelly construction
The jelly is constructed without any implication of conscience and focus.
The jelly construction is based on the LACS universe building approach (LACS, MoLACS, MuLACS, LACS consolidation).
Practically the jelly construction is a universe building process.
The jelly is the set of nodes and the set of links.
Nodes includes concepts, words and variants.
Links are classified according to involved nodes and possible qualification.
The jelly is built through the following steps:
Finding Sources
Corpus are found on the internet for various languages (at least french and english). A corpus is a list of words and variants, with correspondance between them, and it includes the variant forms (for variants) and the lexical category (for words). For the french, the version used has been found on the abu.cnam.fr wes site.
Besides corpus, full text of books are found on the internet and loaded. The source used curently is www.gutenberg.org.
Building base tables
From the corpus material, the following tables are built:
- list of languages
- list of variant forms
- list of lexical categories
- HashSet of u-variants
- HashSet of q-variants
- HashSet of words
- WV links
Building frequency tables
Using the book set frequency tables are built per word node.
Building numeric concept
The next step is the construction of the first concepts.
Numeric concept nodes are first created.
Then numeric connecting concepts are created.
Then the QC links between numeric connecting concepts and numeric concepts are defined.