| المستخلص: |
Each educational institution aims to provide services with high quality and that is because of the high competition and also due to the lack of jobs, that do not fit with the number of the graduate students from Statistical institutes. These services can be through rehabilitation, training and developing of the graduates, monitoring and evaluating the outcomes of plans. Data mining techniques is considered as one of the important methods that give indications that contribute to the performance of organizations and then to the quality of their outcomes. This research presents a practical study in the field of Knowledge Discovery from Data (KDD) using Data Mining techniques; also the research aims to discover some Patterns in the students' academic data. This is achieved by considering the students' academic records that come from Faculty of Technology of Mathematical Sciences and Statistics. The research thereafter obtains general indicators about the academic performance of students and departments in the Faculty, and then finds relationships between courses and their impacts on the graduation results as well as analyzing the input data with respect to gender type. All these have been done via several tasks provided by data mining techniques; such as Classification, Clustering and the Association Rules. To achieve the objectives of this research, we used Classification Algorithms to build predictive models to help decision-makers to build plans that contribute to improve the academic performance of students. The clustering algorithms are then used to cluster graduate students into clusters to give clues about the performance of different departments in the considered faculty. As well, we measured the impacts of the previous records, investigating the relationship between them by using the Association Rules Algorithms. The algorithms have been fitted to a database that acquired from multi sets graduate student records of the Faculty of Technology of Mathematical Sciences and Statistics based on the time period between 2008 and 2013. The analysis has been implemented using the WEKA and the R programming packages. These packages are open sources that most widely used in data mining discipline. Through the indicators we suggested some methods that may help the decision-makers to improve the academic performance and consequently expand employment opportunities for the Faculty and the University graduates.
|
