Application of Genetic Algorithm in Modeling University Admission Decision Support System
Chapter One
Objective of the Study
The main objective of this thesis is to design a model using genetic algorithm that can be employed in searching trends or pattern in student’s previous admission records. This is achieved by using the aggregation of UME, Post UME, O’level scores against their corresponding CGPA at the end of their first academic year in the University. Realizing this objective can help in candidates’ selection criteria for admission process into the university. This main goal can be achieved by means of the following objectives which are:
- To determine the means by which data collected can be translated to meaningful ones.
- To develop a model for searching hidden pattern among the available data set using genetic
- To implement model of genetic algorithm
- To test and validate the model using real data of students’
CHAPTER TWO
REVIEW OF LITERATURE
Introduction
In this chapter, a review is made of predictive technique in genetic algorithm. A description of the research that has been accomplished in the area of university admission decision making is also provided.
History of Evolutionary Algorithms
In the 1950’s and 1960’s several computer scientists independently studied evolutionary systems with the idea that evolution could be used as an optimization tool for engineering problems. The idea in all these systems was to evolve a population of candidate solutions to a given problem, using operators inspired by natural genetic variation and natural selection.
In the 1960s, Rechenberg introduced “evolution strategies” as a method he used to optimize real valued parameters for devices such as airfoils. This idea was further developed by (Schwefel, 1977). The field of evolution strategies has remained an active area of research, mostly developing independently from the field of genetic algorithm (although recently the two communities have begun to interact). Fogel et al (1966) developed “evolutionary programming,” a technique in which candidate solutions to given tasks were represented as finite state machines, which were evolved by randomly mutating their state transition diagrams and selecting the fittest. Together, evolution strategies, evolutionary programming, and genetic algorithms form the backbone of the field of evolutionary computation.
Several researchers working in the 1950s and the 1960s developed evolution inspired algorithms for optimization and machine learning. Box (1957), Friedman (1959), Bledsoe (1961), Bremermann (1962), and Reed et al (1967) all worked in this area, though their work has been given little or none of the kind of attention or follow up that evolution strategies, evolutionary programming, and genetic algorithm have seen. In addition, a number of evolutionary biologists used computers to simulate evolution for the purpose of controlled experiments (Baricelli 1957, 1962; Fraser 1957; Martin and Cockerham 1960). Evolutionary computation was definitely in the air in the formative days of the electronic computer.
Genetic algorithm (GA) was invented by John Holland and his students and colleagues at the University of Michigan in the 1960s and the 1970s (Holland, 1975). In contrast with evolution strategies and evolutionary programming, Holland’s original goal was not to design algorithm to solve specific problems, but rather to formally study the phenomenon of adaptation as it occurs in nature and to develop ways in which the mechanisms of natural adaptation might be imported into computer systems.
Holland’s 1975 book “Adaptation in Natural and Artificial Systems” presented the genetic algorithm as an abstraction of biological evolution and gave a theoretical framework for adaptation under the GA. Holland’s GA is a method for moving from one population of “chromosomes” (for example, strings of ones and zeros, or “bits”) to a new population by using a kind of “natural selection” together with the genetics inspired operators of crossover, mutation, and inversion. Each chromosome consists of “genes” (for example, bits), each gene being an instance of a particular “allele” (for example, 0 or 1). The selection operator chooses those chromosomes in the population that will be allowed to reproduce, and on average the fitter chromosomes produce more offspring than the less fit ones. Crossover exchanges subparts of two chromosomes, roughly mimicking biological recombination between two single chromosome (“haploid”) organisms; mutation randomly changes the allele values of some locations in the chromosome; and inversion reverses the order of a contiguous section of the chromosome, thus rearranging the order in which genes are arrayed.
Holland’s introduction of a population based algorithm with crossover, inversion, and mutation was a major innovation. Unlike Rechenberg’s evolution strategies which started with a “population” of two individuals, one parent and one offspring, the offspring being a mutated version of the parent; many individual populations and crossover were not incorporated until later. (Fogel et al., 1966) evolutionary programming likewise used only mutation to provide variation. Moreover, Holland was the first to attempt to put computational evolution on a firm theoretical footing (Holland, 1975). Until recently this theoretical foundation, based on the notion of “schemas,” was the basis of almost all subsequent theoretical work on genetic algorithms.
CHAPTER THREE
MODELING THE STUDENT’S PERFORMANCE DECISION SUPPORT SYSTEM
System Description
A predictive model which determines the relationships between the entry requirement and the first year CGPA is developed. The admission dataset that represents the entry results of student previously admitted into the programme were first normalized and their features extracted. A list of classifiers was generated for each student performance and grouped according to their respective CGPA. Genetic algorithm is then used as an optimization for searching meaningful relationship between the classifiers by means of reproduction till the best combination is found. Figure 3.1 shows the relationship between the system modules.
CHAPTER FOUR
IMPLEMENTATION OF THE DECISION SUPPORT SYSTEM
Introduction
This chapter presents the detailed implementation of the proposed system using the concept of Genetic Algorithm. The students’ admission data set for computer science students, mathematics department of Ahmadu Bello University, Zaria was used in evaluation of the system and the actual implementation is presented in this chapter.
The system was implemented using Java. Java is known to be platform independent, that is, it can be deployed on different platforms (like Microsoft windows, Unix, Mac OS, Android). It has rich libraries to perform any task. It also makes possible to use java classes to handle tables as required by the work.
CHAPTER FIVE
SUMMARY, CONCLUSION AND RECOMMENDATION
Summary
Educational data mining is used to study the data available in the educational field to bring out the hidden knowledge from it. Over the years, several techniques have been used to analyze students’ performance from different points of view. This thesis presents data mining in education environment that identifies students’ best patterns for admission into the University using Genetic Algorithm technique. The evaluation was done using undergraduate students’ admission results into computer science, department of Mathematics, Ahmadu Bello University. The model was developed based on some selected input variables from the pre-admission data of three different session of students’ offered admission into computer programme. Three examinations (UTME, PUTME and O’level) were used to group the students. The combination of these multiple examinations leads to a significant accuracy to support the decision. This thesis established a performance patterns for the first year entry results of students’ in the University. The findings reveal that Mathematics and Physics have prominent features among the best pattern for almost all the selection.
Conclusion
This thesis used searching technique that identifies students’ best patterns for admission into the University using Genetic Algorithm technique. The result obtained from the analysis of students’ academic performance corroborate most of the literature assertions, thus, it showed that the Genetic algorithm application was able to model the data set used. It can be concluded from the result of the analysis that the current attention that is given to mathematics and physics as one of the basic requirement into the department as a variable in the selection process is a step in the right direction.
Recommendations
This thesis shows that Genetic Algorithm has the capability to bring out hidden knowledge from admission system. Hence, the searching can be giving more strength if all the factors that constitute to student academic performance are provided.
The limitation of this model stems from the fact that not all the relevant factors that play key roles in student performance were available to the researcher. The research was based on academic achievement only, factors like student background, age, number of years spent before getting admission, motivation, financial aids, and so on were not at the researcher disposal. Adding some of these factors will no doubt have influence on the pattern.
An extension of this research should be made to other departments. The current admissions system should be reviewed in order to improve the standard of candidates being admitted into the institution. Genetic algorithm with other machine learning tool should be combine which will be very useful for such an exercise.
REFERENCES
- Adeniyi, O.S., Araoye, M.A., Amali, E. O., Eru, E.U., Ojabo, C. O., and Alao, O.O. (2010). Effect of Using Combination of O’level Result with JAMB Score on Student Performance in the First Two Years of Medical School in Benue State University, Makurdi. Afr. J. Biomed. 13( 3), 189 – 195. Retrieved from http://www.bioline.org.br/request?md10031. Accessed on 30/11/2012.
- Adewale, O., Adebiyi, A., and Solanke, O. (2007). Web-Based Neural Network Model for University Undergraduate Admission Selection and Placement.Pacific Journal of Science and Technology, 8(2), 367-384. Retrieved from http://www.akamaiuniversity.us/PJST8_2_367.pdf. Accessed on
- Adeyemi, T. O. (2010). Credit in Mathematics in Senior Secondary Certificate Examinations as a Predictor of Success in Educational Management in Universities in Ekiti and Ondo States, Nigeria. Middle-East Journal of Scientific Research, 5 (4), 235-244. Retrieved from http://www.idosi.org/mejsr/mejsr5%284%29/8.pdf. Accessed on 22/8/2013.
- Adeyemo, A. B., and Kuye, G. (2006). Mining Students’ Academic Performance using Decision Tree Algorithms. Journal of Information Technology Impact, 6(3), 161-170. Retrieved from http://www.jiti.com/v06/jiti.v6n3.161-170.pdf
- Aigbe, P.E. (2008). Design and Implementation of a Web Based University Admission and Placement Neural Network Model. Unpublished M.Sc. Thesis. Ahmadu Bello University Zaria, Kaduna State.
- Ajala, O. P. (2010). Three Years of Post UME Screening: Influence on Science Education Students’ Achievement. Delta State University, Abraka. International Journal Education Science, 2(1), 29-40. Retrieved from
- http://www.krepublishers.com/02-Journals/IJES/IJES-02-0-000-10-Web/IJES 02-1000-10-Abst-PDF/IJES-02-1-29-10-031-Patrick-A-O/IJES-02-1-29-10-031
- Patrick-A-O-Tt.pdf