Information Technology Reference
In-Depth Information
2.4.2.1 Examples of Operation
The above described rule-based fuzzy logic system was used in a postgradu-
ate program in the field of informatics at the University of Piraeus in Greece. It
was used in order to offer dynamically personalized e-training in computer pro-
gramming and the language C. At the beginning, all the domain concepts of the
learning material were considered to be 'Unknown' for the learners. At the next
interactions, the system delivered to them the appropriate learning material for
each individual student's needs by adapting instantly to the learner's individ-
ual learning pace. The KL value of each domain concept was determined by the
results of the tests. There were two kinds of tests: (i) the tests that corresponded
to each individual domain concept of the learning material (practice tests), (ii) the
final tests that corresponded to the sections of the learning material (they included
exercises of a variety of domain concepts). In particular, each time the learner read
a domain concept, s/he had to complete a corresponding practice test. When, the
learner had completed successfully all the practice tests of the domain concepts of
a section (e.g. iterations with concrete number of loops, arrays, sub-programming),
then s/he had to complete the final test of the section. If s/he succeeded to the final
test, then s/he transited to a next section. Otherwise, s/he had advised to revise
some domain concepts. Representative examples of the system's implementation
follow.
Example 1
George had learned the sections 1 (domain concepts 1.1 to 1.7) and 2 (domain
concept 2.1) and she was taught the domain concepts of the section 3 (domain
concepts 3.1 to 3.3) (Interaction I of Table 2.8 ). He read the concept C 3.1 . Then,
he was examined in the particular domain concept and succeeded 78 %. According
to the above, the value of the defined membership functions for concept C 3.1
become μ Un = 0, μ MKn = 0, μ Kn = 0.4, μ L = 0.6 and μ A = 0. According to the
FR-CN (Fig. 2.11 ) the concept C 3.1 affects the following concepts C 3.2 and C 3.3
with “strength of impact” 0.5 and 0.2 correspondingly. Consequently, applying the
fuzzy rule R2 (b) and (c), KL(C 3.2 ) becomes 20 % 'Known' and 30 % 'Learned'.
The rest 50 % of the particular concept remains 'Unknown' (Interaction II of
Table 2.8 ). Similarly, applying the same rules, KL(C 3.3 ) becomes 8 % 'Known'
and 12 % 'Learned'. The rest 80 % of the particular concept remains 'Unknown'
(Interaction II of Table 2.8 ). Therefore, although concepts C 3.2 and C 3.3 are not
completely unknown to George, the system advises him to read them.
Example 2
Kate had learned the sections 1 (domain concepts 1.1 to 1.7), 2 (domain con-
cept 2.1), 3 (domain concepts (3.1 to 3.3) and the concepts 4.1, 4.5 and 5.5
(Interaction I of Table 2.9 ). She read the concept C 4.2 to improve her knowledge
level. Then, she was examined in the particular domain concept and succeeded
86 %. According to the above, the value of the defined membership functions
for concept C 4.2 become μ Un = 0, μ MKn = 0, μ Kn = 0, μ L = 0.8 and μ A = 0.2.
Search WWH ::




Custom Search