Information Technology Reference
In-Depth Information
Fig. 3. Discrete Bayesian Network Modelling for the Diagnosis of Deterioration of
Semantic Content
Where:
-
α
is a normalising constant.
P
(
sex,Alzheimer
)
P
(
Alzheimer
)
-
P
(
sex
|
Alzheimer
)=
P
(
age,Alzheimer
)
P
(
Alzheimer
)
-
P
(
age
|
Alzheimer
)=
P
(
EducationalLevel,,Alzheimer
)
P
(
Alzheimer
)
P
(
EducationalLevel
|
Alzheimer
)=
-
For Cognitive Deterioration the same method can be used. Furthermore, an
automatic learning algorithm can also be used for the intermediate variables and
variables of interest. The total attribute production for each semantic category
of each of the objects only has to be added. Then the K-Means algorithm is
applied (using WEKA, fig ) to create a specific number of data clusters, using
the Euclidean distance for this. The K-Means algorithm is based on the patient
producing attributes to determine the centroids of each cluster. Each cluster
represents each of the states of the variable. Once the centroids of the different
clusters have been defined, i.e. the different levels of cognitive deterioration, the
database is analysed to calculate the conditional probabilities using the following
formulation:
N
(
DCx
1
,DCSV x
2
,EAx
3
)
N
(
DCx
2
,EAx
3
)
P
(
DCx
1
|
DCSV x
2
,FAx
2
)=
where
Search WWH ::
Custom Search