Exploiting the Core Academic Performance Prediction Parameters Using a Neuro-Fuzzy Model

Journal Title: American journal of Engineering Research - Year 2017, Vol 6, Issue 6

Abstract

This work developed a Neuro-Fuzzy model as an intelligent computational framework to provide decision support for the admission of candidates into Higher Institutions. The traditional admission process is based on Unified Tertiary Matriculation Examination (UTME) and Post- Unified Tertiary Matriculation Examination (PostUTME) scores, so long as the O-Level requirements are satisfied. This method has proven to have limitations. In this work, a five layer Adaptive Neuro-fuzzy Inference System (ANFIS) is modeled using a fuzzy logic input decision variables, taking into consideration the aggregate previous academic performances of the candidate and other related parameters that can influence academic performance. Historical records of students’ academic performances are used to build the fuzzy logic decision tree from which the initial fuzzy logic rules are extracted. The MATLAB NeuroFuzzy Designer is used for the modeling and training of the ANFIS model. Validation of the accuracy of the prediction using the affected student’s year 1, 2 , 3, 4 and 5 CGPAs, give the coefficients of multi determination to be approximately 0.96, 0.97 , 0.99, 0.98 and 0.96 respectively. These results show very high degree (97.2%) of accuracy, between the predicted and actual performance. These findings show that the developed model can be relied upon to make decisions on the admission of candidates into HAILs in any country of the world.

Authors and Affiliations

Okereke Eze Aru,, Nkwachukwu Chukwuchekwa,, F. K. Opara

Keywords

Related Articles

Analysis of Improvement in Aerodynamics of a Vehicle Type Baja SAE - And the Impact on Performance

The constant search for greater efficiency in any type of project, is one of the top searches of a professional in engineering, this being linked directly to cost reduction, performance improvement and analysis from an e...

Aid and Assistance China for Development Socio-Economical and Cultural To Central African Republic

The Central African Republic like the heart World, through the vicissitudes given its geographical position. Its accession in 1960 to the Maoist movement has been consolidated by the ChinaAfrica Forum held in October 200...

Non Newtonian Behavior of Blood in Presence of Arterial Occlusion

The objective of the present numerical model is to investigate the effect of shape of stenosis on blood flow through an artery using Bingham plastic fluid model. Blood is modeled as Bingham plastic fluid in a uniform cir...

Corrosion Behavior of Buried Carbon Steel Pipeline In Iraqi Soil in Different Sodium Chloride Concentrations

In this work, external corrosion behavior of oil pipeline was evaluated using potentiodynamic polarization curves. The corroding metal was carbon steel. To estimate corrosion by potential and current density of carbon st...

Using Cryptography Algorithms to Secure Cloud Computing Data and Services

These days regrading to the high demand on usingthe cloud computing services for storing and processing data, there is awareness about the information security and cloud computing. This paper present and take you to see...

Download PDF file
  • EP ID EP401423
  • DOI -
  • Views 43
  • Downloads 0

How To Cite

Okereke Eze Aru, , Nkwachukwu Chukwuchekwa, , F. K. Opara (2017). Exploiting the Core Academic Performance Prediction Parameters Using a Neuro-Fuzzy Model. American journal of Engineering Research, 6(6), 232-245. https://europub.co.uk/articles/-A-401423