Information Technology Reference
In-Depth Information
higher than those three weeks ago. If they are only at least 30% higher, it is assessed
as "sharp increase", and if they are only at least 5% higher, it just an "increase".
Weekly Data (Incidences)
1st Week
2nd Week
3rd Week
4th Week
Temporal Abstraction
Course Description Parameters
3 Trend Assessments + 4 weekly Incidences
Re trie val: Distanc es
List of All Former Courses
sorted according distances to query course
Sufficient Similarity
List of Most Similar Former Courses
sorted according distances to query course
Adaptation
Warning if appropriate
Fig. 3. The prognostic model for TeCoMed
Together with the four weekly data these assessments are used to determine
similarities between a query course and all courses stored in the case base. Our
intention for using these two sorts of parameters is to ensure that a query course and
an appropriate similar course are on the same level (similar weekly data) and that they
have similar changes on time (similar assessments).
3.2
Searching for Similar Courses
So far, we compute distances between a query course and all courses stored in the
case base sequentially. In the future we hope to develop a more efficient retrieval
algorithm. The considered attributes are the three nominal valued trend assessments
and the four weekly incidences. The similarity measure is only based on our
 
Search WWH ::




Custom Search