The other day there was a discussion that why some questions are repeatedly asked in the question papers. For example if a question is to be asked from clustering topic in data mining subject, it is usually on k-means clustering. Similarly in AR mining, it is Apriori algorithm. Certain examples / algorithms become so popular because of their simplicity and ease of understanding that they become most suitable tool for explaining that topic. They are exemplars in that topic. So if a question is to be asked from that, it is always the exemplar which is asked.
How to take care of such exemplars while mapping the QP to syllabus ontology to find the syllabus coverage? After extracting the concepts from the questions we are trying to map those to the syllabus ontology and colouring all those nodes covered by the question. Suppose under the clustering node in the syllabus ontology, there are many other nodes. But only the k-means clustering node gets coloured always. Should we consider that the clustering node is covered completely even if only one node from that is coloured. I think we can if that particular node is an exemplar of that topic.
It is very true that some things which come so naturally to humans become very difficult when it has to be automated.
No comments:
Post a Comment