METHODOLOGICAL IMPLEMENTATION OF ACADEMIC ORGANIZATIONAL OPERATED DATA WITH CLUSTERING MECHANISM USING AN IMPROVED KMEANS ALGORITHM
Keywords:
K-means; density; outlier; initial cluster centers; college student;Abstract
Many colleges have accumulated a large amount of
information, such as achievement data and consumption records.
According to the above information, we attempt to identify the
student group from various aspects. Given this, we can acquire
the characteristics of students in different groups. In this way,
the college can have a better understanding of students to
accomplish the reasonable management. To obtain more
accurate cluster results, we proposed an improved K-means
algorithm. Specially, we effectively detect outliers based on the
grid density. In addition, we design a new method to produce
initial cluster centers which replaces the traditional random way.
Real experiments are conducted and the results show the
iteration time is reduced and clustering precision is improved
