Information Technology Reference
In-Depth Information
need to be traversed if the parent tree also has the key value; and it is more
important to note the types of structure and links that get created, rather than the
exact semantics of the whole process. For example, start with the
first linking
keyset, remove a link key if it gets used and only continue if the next link key is in
the list.
Not shown in the
figure, the key sets might possibly be similar to the NoSql
column-family stores, with a primary key retrieving a set of secondary keys. The
secondary keyset would then allow for the traversal of the linked database trees.
The primary key can relate to an external entity that is interested in certain groups
of concepts. As each tree is quite separate, the secondary keys could, for example,
relate or point to sets of base tree concepts. In this case, graph-like navigation is
also possible, as leaf nodes can link to other tree base nodes as well, where the
secondary keyset helps to de
ne the allowed starting paths.
4.2.1 Query Example
A query process is still required to access the tree contents and can be related to the
keysets that exist. For example, if the cat also starts to eat grass, a primary key that
points to the
'
'
'
'
tree in its secondary
keyset as well. The ' Cat ' link now also includes the G graph link. Without any
additional information, the traversal can return either the cat or the elephant for milk
and grass. This might need to be speci
Thirsty
tree, might then include the
Black
ed, because another similarly indexed
primary key could accept, cat, boy or elephant for just the milk concept. It might
also be interesting to consider that the secondary keyset allows for the completion
of circuits through the graph trees. For example, if the
first primary key also
includes the
) in its secondary keyset and requires all
indexed trees to be true; then only cat and elephant can be returned. This is just a
possibility and is related to suggestions about intelligence in Sect. 7.2 . It does
however look like useful searches might require a set of conditions, but also allow
for some automatic reasoning, where a query language is possible. For this
example, it looks like a Horn clause could be used (Jarke et al. 1995 ; Greer 2011 ,
for example).
'
Ate
'
tree (and maybe
'
Drank
'
4.3 Compound Counts
Another requirement is to break the structure up again. If the concept base, for
example, received concept sequences of
'
'
, with no
relation to any of the animal tree base concepts; then the process should just add
'
drank milk with a long trunk
'
'
'
long trunk
to the
drank milk
base. This is not necessarily incorrect because that
speci
c information does not state that the cat or the boy do not have long trunks. If
the information is then used and the cat or boy trees traversed, it will incorrectly
return that they have long trunks. Some type of compound count can be used to
Search WWH ::




Custom Search