Main Article Content
Abstract
Allocating new students into their classes is a clustering problem, that is how to cluster new students into
their classes so that each class contains students in the number that less than or equals to its capacity and has
minimum gap of intelligence. It needs a suitable method to avoid an educational problem. This paper describes
the comparison of Genetic Algorithm (GA) and Modification of Agglomerative Methods (AM) for solving this
problem. To determine which method is better then the other, the software of each method which can cluster n
students with m attributes into c classes are evaluated by two-dimensional random data consists of 200 students.
Then we compare the results. Comparison of GA and AM for clustering the data sets shows that although the GA
cluster the data successfully, the method provides no advantages over AM. Intelligence gap of students in each
class clustered by GA almost same each other, but the average of this value is greater than by AM. Meanwhile,
the intelligence gap of student clustered by AM depend on the clustering sequence. This GA performance may be
is caused by unsuitable GA approach, both chromosome representation and GA operators in this research.
Better GA approach may enhance the effectiveness of the GA searching.
Keywords: Agglomerative Method, cluster, Genetic Algorithm, student.
their classes so that each class contains students in the number that less than or equals to its capacity and has
minimum gap of intelligence. It needs a suitable method to avoid an educational problem. This paper describes
the comparison of Genetic Algorithm (GA) and Modification of Agglomerative Methods (AM) for solving this
problem. To determine which method is better then the other, the software of each method which can cluster n
students with m attributes into c classes are evaluated by two-dimensional random data consists of 200 students.
Then we compare the results. Comparison of GA and AM for clustering the data sets shows that although the GA
cluster the data successfully, the method provides no advantages over AM. Intelligence gap of students in each
class clustered by GA almost same each other, but the average of this value is greater than by AM. Meanwhile,
the intelligence gap of student clustered by AM depend on the clustering sequence. This GA performance may be
is caused by unsuitable GA approach, both chromosome representation and GA operators in this research.
Better GA approach may enhance the effectiveness of the GA searching.
Keywords: Agglomerative Method, cluster, Genetic Algorithm, student.