Article Preview
Top1. Introduction
Education Data Mining (EDM), which is considered a learning science and a solution for continuous growing of data in the education domain (Calders, 2011) improves the managerial decision making of several educational institutions. Large data can be found in these organizations and can be used to easily identify the target tertiary institutions. The motivation of this research is to develop a marketing function that serves as a tool in determining the possible schools for recruitment. The secondary schools are the marketing indicators. This helps the administration of tertiary institutions identify which among the secondary schools do not contribute to the enrollment rate or perhaps no enrollees at all. In effect, it alleviates the expenses of the tertiary institutions.
1.1. Background of the Study
iRecruit function uses a classification technique (Han & Kamber, 2006) in mining data since it is a method that focuses on prediction. It classifies probable institutions for enrollment considering the General Weighted Average (GWA), Income (IncomeBracket), Distance (RadialDistance) and Category of school (SchoolOwnership). Among the different classification techniques, Ross Quinlan’s C4.5 algorithm was used. It is the successor of ID3 (Iterative Dichotomiser 3) algorithm, which is one of the techniques in generating DTs. It is a method that presents the data in a more straightforward interpretation (Jantan, 2010).
The function was written in PERL (Practical Extraction and Report Language). It was developed by Larry Wall in 1987. Perl fills the gap between the capabilities of shell scripts and compiled C programs (Hudson & Hudson, 2007). This scripting language was used due to its expertise in extracting data from a text file and its capability to analyze the decision tree generated by C4.5 algorithm. MySql (Gosselin, 2011) is the database management software used and the application runs in the Ubuntu (Hill, 2007) platform.
A number of studies have been conducted in classification technique, prediction and C4.5.
Ranjan and Malik (2007) developed an effective educational framework using data mining technique to uncover hidden trends and patterns that can be used to produce accurate based prediction in the process of counseling students. The work of Ranjan and Malik only proved that outsized educational data may improve the student services of a particular institution. DT model was used as a framework to classify the historical data of the examinees and enrollees from the different secondary institutions.
Al-Radaideh (2011) used a classification technique of generating decision trees in predicting the suitable track for students in secondary education. The work of Al-Radaideh verifies that classification method of data mining supports the academic track planning of the students. The research implemented a classification technique to help the marketing department of every educational organization identify which among the high school institutions need to be visited.
Dekker (2009) used classifiers in a form of decision trees in predicting the possible drop out of EE students and concluded that the accuracy result of these classifiers is between 75 to 80%. The work of Dekker proved that the classification technique can be used as a reliable forecast. In view of this, the classifiers are the qualified attributes as specified by the user. These attributes served as the marketing criteria in choosing the potential schools to be visited.
2. Irecruit Function
This section confers the integration of C4.5 algorithm in generating the potential institutions for recruitment.