Abstract— Higher Education is a level of education after
secondary education which includes diploma programs,
undergraduate programs, master programs, doctoral programs,
professional programs, and specialist programs organized based
on the culture of the Indonesian nation. Student graduation is one
of the important factors to improve university accreditation.
Students who graduate above 5 years and the number of students
who drop out are important indicators in determining
accreditation which then causes the difficulty of accrediting a
college to rise. This research aims as an early warning for students
who graduate on time and graduate late from the Faculty of
Information Technology, Institute of Science and Business Atma
Luhur using the C4.5 decision tree algorithm by implementing the
Cross-Industry Standard Process for Data Mining (CRISP- DM)
method. The initial data of this research amounted to 1,015 which
was taken through a query in the database of the Atma Luhur
Institute of Science and Business. However, the data that will be
used becomes 694 after preprocessing due to the large number of
record contents that do not have a graduation year, with a total of
641 graduates graduating on time and 53 graduates graduating
late. Based on the application of the model using the C4.5 decision
tree algorithm and the Confusion Matrix method, the accuracy is
93.94%, Recall is 98.59%, and Precision is 95.03%. So it can be
concluded that the C4.5 decision tree algorithm is the most
effective algorithm for predicting student graduation, because it
has a high level of accuracy.