Information Technology Reference
In-Depth Information
Tabl e 1 . Abstract of Database of Instances (Linguistic Corpus)
PK category taxonomic types parts functional evaluative place conduct cause procedural life cycle other
1
car
0
1
5
5
0
0
0
0
0
0
3
1
apple
1
2
0
6
0
0
0
1
0
0
0
1 trousers
0
8
0
2
0
0
0
0
0
0
0
1
dog
0
6
0
1
4
0
0
0
0
0
1
1
pine
0
0
0
5
0
0
0
0
0
0
0
1
chair
0
4
0
2
0
3
0
0
0
0
0
...
...
...
...
...
...
...
...
...
...
...
...
...
This work starts from a set of data obtained from the works by [12][14], where
semantic category definition features of the linguistic Corpus of healthy subjects
and subjects with Alzheimer's Disease are analysed. In other words, we start with
a base of instances (table 1 shows an abstract) where the patients' definitions are
classified into 11 basic conceptual blocks: taxonomical, types, parts, functional,
evaluative, place/habitat, behaviour, causes/generates, procedural, life cycle and
others. In other words, each patient's definition is analysed and the number of
attributes that he produces is classified into these 11 basic conceptual blocks,
with two differential semantic categories, living creatures and artefacts. The
following table shows an abstract of the linguistic corpus (database of instances)
that this research project has started from.
The research work is in its initial stage. Therefore, this article only describes
the model for the diagnosis and the decisions that have been taken for its design.
We think that in the very near future we can obtain promising results, especially
when we have a wider database of instances, which will enable us to do su
ciently
reliable and assessable experiments.
2
Justification for the Technique
Bayesian Networks and influence diagrams offer a number of advantages that
make them attractive for use in this application field. The advantages can be
summarised in [11,6]:
-
Bayesian Networks are Soft Computing techniques and they have to be used
because there is currently no deterministic method for diagnosing Alzheimer.
-
Bayesian Networks are based on compact graphs and are intuitive of a causal
relation between entities of a problem in a specific domain. Other techniques
such as Neuronal Networks or Bioinspired Techniques are based on graph
representations that are dicult for experts in the field to read.
-
Inference is based on the theory of calculating probabilities and the decision
theory. It therefore provides a coherent mathematical method to derive con-
clusions in uncertainty, where multiple sources of information are involved
in complex interaction patterns. Other soft computing techniques are also
based on complex mathematical models, where the inference explanation is
very complex.
Search WWH ::
Custom Search