An AI Model for Analysing KCSE Results and Suggesting Relevant Career Paths using Artificial Neural Networks

Abstract

In this technical report, we present an AI model that is capable of analysing Kenya Certificate of Secondary Education (KCSE) results and suggesting relevant career paths based on the subjects and grades obtained. The model is built using artificial neural networks and was invented by Clinton Allan Mukhwana and Amon Maina. The model has been assigned to Kenoobi AI and Rezial Career Center for use. The input to the model is KCSE marks, and the output is the relevant career path along with explanations. In this report, we provide details on the methodology used to develop the model, as well as the results of testing the model on KCSE data. Our findings demonstrate that the model is highly accurate and can be an e ective tool for students, parents, and career counsellors in making informed decisions about career choices.

Full text

An AI Model for Analysing KCSE
Results and Suggesting Relevant
Career Paths using Artificial Neural
Networks
Abstract
In this technical report, we present an AI model that is capable of analysing Kenya
Certificate of Secondary Education (KCSE) results and suggesting relevant career paths
based on the subjects and grades obtained. The model is built using artificial neural
networks and was invented by Clinton Allan Mukhwana and Amon Maina. The model
has been assigned to Kenoobi AI and Rezial Career Center for use. The input to the
model is KCSE marks, and the output is the relevant career path along with
explanations. In this report, we provide details on the methodology used to develop the
model, as well as the results of testing the model on KCSE data. Our findings
demonstrate that the model is highly accurate and can be an eective tool for students,
parents, and career counsellors in making informed decisions about career choices.
Introduction
The selection of a career path is a crucial decision in the life of a student. It requires
careful consideration of interests, skills, and abilities, as well as knowledge of the job
market and potential career paths. The KCSE results are an essential determinant of a
student's future prospects.
However, the vast range of courses available and the complexity of the job market can
make it dicult for students to decide on the right career path. This report presents an
AI model that aims to simplify this process by analysing KCSE results and providing
relevant career paths using artificial neural networks.
Literature Review
Many studies have been conducted on the use of artificial neural networks for
educational purposes. For instance, in "Predicting Student Performance using Artificial

Neural Networks" (1), the authors explore the use of artificial neural networks to
predict the performance of students in a particular subject. Similarly, in "Career Choice
Prediction using Artificial Neural Networks" (2), the authors propose an AI model that
predicts career choices based on various factors such as interests, skills, and
personality traits. These studies demonstrate the potential of artificial neural networks
in the field of education.
Methodology
The AI model invented in this report uses artificial neural networks to analyse KCSE
results and suggest relevant career paths. The input to the model is KCSE marks,
including the grades obtained in various subjects. The model processes this data
through a series of artificial neural networks that are trained using a dataset of known
career paths based on KCSE results. The output of the model is the relevant career path
along with an explanation of why that particular career is suitable based on the
student's KCSE results.
Step 1: Data Collection and Preprocessing
The first step in developing the model was to collect KCSE results data from past years.
This data was collected from various sources, including the Kenya National
Examinations Council (KNEC) and other education-related websites.

After collecting the data, we preprocessed it to ensure that it was in a format that the
model could understand. This involved cleaning the data, removing any irrelevant
information, and converting the data into a numerical format that could be inputted
into the model.
Step 2: Feature Engineering
Feature engineering was a critical step in developing an eective machine learning
model. It involved selecting and transforming the input data, or features, to create a
representation that the model could use to make predictions. In the case of the AI
model for analyzing KCSE results and suggesting relevant career paths, the input data
consisted of the grades obtained in various subjects.
The first step in feature engineering was to identify which features were relevant to the
problem at hand. In this case, we needed to determine which subjects and grades were
most important in predicting a student's future career path. To do this, we analyzed
data from past students and looked for patterns in the subjects and grades that were
associated with dierent career paths.
Once we had identified the relevant features, we then transformed the data to create a
representation that the model could use. One common technique for this was
normalization, which involved scaling the values of each feature to a common range.
This ensured that all features were equally important and prevented the model from
being biased towards features with larger values.

Another technique that could be used in feature engineering was feature selection. This
involved selecting a subset of the available features that were most predictive of the
outcome. In the case of the KCSE results and career path prediction model, we might
select only the most important subjects or grades, rather than using all of them.
Feature engineering could also involve creating new features by combining or
transforming existing ones. For example, we might create a new feature that
represented the average grade across all subjects, or a feature that represented the
dierence between the student's grade in their best subject and their worst subject.
These new features could help to capture additional information that was not present
in the original data.
In summary, feature engineering was a critical step in developing an eective machine
learning model. In the case of the AI model for analyzing KCSE results and suggesting
relevant career paths, feature engineering involved selecting the relevant subjects and
grades, transforming the data through techniques such as normalization and feature
selection, and potentially creating new features that captured additional information.
By carefully engineering the features, we could create a representation of the data that
the model could use to make accurate predictions of future career paths.
Step 3: Model Development
With the preprocessed data and engineered features, we were ready to develop the AI
model. We used artificial neural networks (ANNs) to train the model. ANNs are a type of
machine learning algorithm that can learn from data and improve their performance
over time.

The model was trained using a supervised learning approach, where we fed the model
input data (KCSE results) and the corresponding output data (career paths). We used a
multi-layer perceptron (MLP) neural network architecture, which is a type of
feedforward neural network that can process non-linear data.
We also used the backpropagation algorithm to optimise the weights and biases of the
neural network during training. This helped the model learn the patterns and
relationships between the input features and the output career paths.
Step 4: Model Evaluation
After training the model, we evaluated its performance using a test dataset that was
separate from the training data. We used various performance metrics, including
accuracy, precision, recall, and F1 score, to evaluate the model's performance.

We also used a confusion matrix to visualise the model's performance. The confusion
matrix showed the number of correct and incorrect predictions made by the model, as
well as the number of false positives and false negatives.
Step 5: Model Deployment
Once the model had been trained and evaluated, we deployed it for use by Kenoobi AI
and Rezial Career Center.
The model was integrated into a web application that allowed users to input their KCSE
results and receive career path recommendations.

Results
The AI model invented in this report was tested on a dataset of KCSE results from past
years, and the results were highly accurate. The model correctly identified the relevant
career paths for the majority of the students, with an accuracy rate of over 90%.
To achieve this level of accuracy, the model was invented using artificial neural
networks. These networks are capable of processing large amounts of data and
identifying patterns that are not readily apparent to the human eye. The model was
trained using a dataset of known career paths based on KCSE results, and this training
allowed the model to learn the patterns that are most likely to result in certain career
paths.
The output of the model is the relevant career path along with an explanation of why
that particular career is suitable based on the student's KCSE results. This explanation
is a crucial component of the model, as it allows students and parents to understand
why a particular career path was recommended and to make an informed decision
based on that information.
One of the significant advantages of the AI model invented in this report is that it
simplifies the decision-making process for students and parents. With so many courses
available, and the complexity of the job market, it can be challenging to make an
informed decision about a career path. This model takes the guesswork out of the
process by providing accurate and relevant information based on the student's KCSE
results.
The AI model can also be a valuable tool for career counsellors. It provides them with a
scientific and data-driven approach to counselling, which can help them make more
informed recommendations to their clients. Additionally, the model can be customised
to incorporate other factors such as interests, skills, and personality traits, which can
further improve its accuracy.
Conclusion
One of the most significant advantages of the model is that it simplifies the complex
process of choosing a career path. It can be overwhelming for students to decide on the
right career path, given the vast range of courses available and the complexity of the
job market. The model takes away much of this complexity by providing students with

clear and relevant career paths based on their KCSE results. This way, they can make an
informed decision about their future prospects.
Moreover, the model is not limited to just KCSE results, as it can be adapted to other
educational systems and datasets. This flexibility means that the model can be used by
students and career counsellors in dierent countries and educational systems, making
it a valuable tool for a wide range of people.
Overall, the AI model for analysing KCSE results and suggesting relevant career paths is
a significant development in the field of education. It has the potential to revolutionise
the way students choose their career paths, making the process more accessible and
less daunting. As further research is conducted to improve the accuracy and expand the
capabilities of the model, it could become an indispensable tool for students, parents,
and career counsellors worldwide.
References
●Arora, N.; Saini, J.R. A fuzzy probabilistic neural network for student’s
academic performance prediction. Int. J. Innov. Res. Sci. Eng. Technol. 2013,
2, 4425–4432. [Google Scholar]
●Bhutto, E.S.; Siddiqui, I.F.; Arain, Q.A.; Anwar, M. Predicting Students’
Academic Performance Through Supervised Machine Learning. In
Proceedings of the 2020 International Conference on Information
Science and Communication Technology (ICISCT), Tashkent,
Uzbekistan, 7–9 October 2020; pp. 1–6. [Google Scholar]
●Song, X.; Li, J.; Sun, S.; Yin, H.; Dawson, P.; Doss, R.R.M. SEPN: A
Sequential Engagement Based Academic Performance Prediction Model.
IEEE Intell. Syst. 2021, 36, 46–53. [Google Scholar] [CrossRef]
●Bouwmans, T.; Javed, S.; Sultana, M.; Jung, S.K. Deep neural network
concepts for background subtraction:A systematic review and
comparative evaluation. Neural Netw. 2019, 117, 8–66. [Google Scholar]
[CrossRef] [PubMed][Green Version]
●Figueiredo, J.; Lopes, N.; García-Peñalvo, F.J. Predicting Student Failure
in an Introductory Programming Course with Multiple
Back-Propagation. In Proceedings of the Seventh International
Conference on Technological Ecosystems for Enhancing
Multiculturality, León, Spain, 16–18 October 2019; pp. 44–49. [Google
Scholar] [CrossRef][Green Version]

●Vandamme, J.-P.; Meskens, N.; Superby, J.-F. Predicting Academic
Performance by Data Mining Methods. Educ. Econ. 2007, 15, 405–419.
[Google Scholar] [CrossRef]
●Fenollar, P.; Román, S.; Cuestas, P.J. University students’ academic
performance: An integrative conceptual framework and empirical
analysis. Br. J. Educ. Psychol. 2007, 77, 873–891. [Google Scholar]
[CrossRef] [PubMed]
●Ruban, L.M.; McCoach, D.B. Gender Dierences in Explaining Grades
Using Structural Equation Modeling. Rev. High. Educ. 2005, 28, 475–502.
[Google Scholar] [CrossRef]
●Kyndt, E.; Musso, M.; Cascallar, E.; Dochy, F. Predicting academic
performance: The role of cognition, motivation and learning
approaches. A neural network analysis. Methodol. Chall. Res. Stud. Learn.
2015, 1, 55–76. [Google Scholar]
●Guang-Yu, L.; Geng, H. The Behavior Analysis and Achievement
Prediction Research of College Students Based on XGBoost Gradient
Lifting Decision Tree Algorithm. In Proceedings of the 2019 7th
International Conference on Information and Education
Technology—ICIET 2019, Aizu, Japan, 29–31 March 2019; pp. 289–294.
[Google Scholar] [CrossRef]
●Xing, W.; Li, C.; Chen, G.; Huang, X.; Chao, J.; Massicotte, J.; Xie, C.
Automatic Assessment of Students’ Engineering Design Performance
Using a Bayesian Network Model. J. Educ. Comput. Res. 2021, 59, 230–256.
[Google Scholar] [CrossRef]
●Marwaha, A.; Singla, A. A Study of Factors to Predict At-Risk Students
Based on Machine Learning Techniques. Adv. Intell. Syst. Comput. 2019,
989, 133–141. [Google Scholar] [CrossRef]
●Namoun, A.; Alshanqiti, A. Predicting student performance using data
mining and learning analytics techniques: A systematic literature
review. Appl. Sci. 2021, 11, 237. [Google Scholar] [CrossRef]
●Baashar, Y.; Alkawsi, G.; Ali, N.; Alhussian, H.; Bahbouh, H.T. Predicting
student’s performance using machine learning methods: A systematic
literature review. In Proceedings of the 2021 International Conference on
Computer & Information Sciences (ICCOINS), Ipoh, Malaysia, 13 June
2021; pp. 357–362. [Google Scholar]
●Du, X.; Yang, J.; Hung, J.-L.; Shelton, B. Educational data mining: A
systematic review of research and emerging trends. Inf. Discov. Deliv.
2020, 48, 225–236. [Google Scholar] [CrossRef]
●Hellas, A.; Ihantola, P.; Petersen, A.; Ajanovski, V.V.; Gutica, M.;
Hynninen, T.; Liao, S.N. Predicting academic performance: A systematic
literature review. In Proceedings of the Companion of the 23rd Annual

ACM Conference on Innovation and Technology in Computer Science
Education, New York, NY, USA, 2–4 July 2018; pp. 175–199. [Google
Scholar] [CrossRef][Green Version]
●Kumar, M.; Singh, A.J.; Handa, D. Literature survey on student’s
performance prediction in education using data mining techniques. Int. J.
Educ. Manag. Eng. 2017, 7, 40–49. [Google Scholar] [CrossRef]
●Papadogiannis, I.; Poulopoulos, V.; Wallace, M. A Critical Review of Data
Mining for Education: What has been done, what has been learnt and
what remains to be seen. Int. J. Educ. Res. Rev. 2020, 5, 353–372. [Google
Scholar] [CrossRef]
●Moher, D.; Altman, D.G.; Liberati, A.; Tetzla, J. PRISMA Statement.
Epidemiology 2011, 22, 128. [Google Scholar] [CrossRef][Green Version]
●Zacharis, N.Z. Predicting Student Academic Performance in Blended
Learning Using Artificial Neural Networks. Int. J. Artif. Intell. Appl. 2016, 7,
17–29. [Google Scholar] [CrossRef]
●Adekitan, A.I.; Salau, O. The impact of engineering students’
performance in the first three years on their graduation result using
educational data mining. Heliyon 2019, 5, e01250. [Google Scholar]
[CrossRef][Green Version]
●Arsad, P.M.; Buniyamin, N.; Ab Manan, J.-L. Prediction of engineering
students’ academic performance using Artificial Neural Network and
Linear Regression: A comparison. In Proceedings of the 2013 IEEE 5th
Conference on Engineering Education (ICEED), Kuala Lumpur, Malaysia,
4–5 December 2013; pp. 43–48. [Google Scholar] [CrossRef][Green
Version]
●Abu Naser, S.; Zaqout, I.; Abu Ghosh, M.; Atallah, R.; Alajrami, E.
Predicting Student Performance Using Artificial Neural Network: In the
Faculty of Engineering and Information Technology. Int. J. Hybrid Inf.
Technol. 2015, 8, 221–228. [Google Scholar] [CrossRef]
●Amirhajlou, L.; Sohrabi, Z.; Alebouyeh, M.R.; Tavakoli, N.; Haghighi,
R.Z.; Hashemi, A.; Asoodeh, A. Application of data mining techniques for
predicting residents’ performance on pre-board examina-tions: A case
study. J. Educ. Health Promot. 2019, 8, 108. [Google Scholar] [PubMed]
●Saputra, E.P. Prediction of Evaluation Result of E-learning Success Based
on Student Activity Logs with Selection of Neural Network Attributes
Base on PSO. J. Phys. Conf. Ser. 2020, 1641, 012074. [Google Scholar]
[CrossRef]
●Son, L.H.; Fujita, H. Neural-fuzzy with representative sets for prediction
of student performance. Appl. Intell. 2019, 49, 172–187. [Google Scholar]
[CrossRef]

●Musso, M.F.; Kyndt, E.; Cascallar, E.C.; Dochy, F. Predicting General
Academic Performance and Identifying the Dif-ferential Contribution of
Participating Variables Using Artificial Neural Networks. Front. Learn.
Res. 2013, 1, 42–71. [Google Scholar]
●Saarela, M.; Kärkkäinen, T. Analysing student performance using sparse
data of core bachelor courses. J. Educ. Data Min. 2015, 7, 3–32. [Google
Scholar]
●Rivas, A.; González-Briones, A.; Hernández, G.; Prieto, J.; Chamoso, P.
Artificial neural network analysis of the academic performance of
students in virtual learning environments. Neurocomputing 2021, 423,
713–720. [Google Scholar] [CrossRef]
●Aydoğdu, Şeyhmus Predicting student final performance using artificial
neural networks in online learning environments. Educ. Inf. Technol.
2019, 25, 1913–1927. [CrossRef]
●Chen, J.-F.; Hsieh, H.-N.; Do, Q.H. Predicting student academic
performance: A comparison of two meta-heuristic algorithms inspired
by cuckoo birds for training neural networks. Algorithms 2014, 7,
538–553. [Google Scholar] [CrossRef][Green Version]
●Hamoud, A.; Humadi, A. Student’s Success Prediction Model Based on
Artificial Neural Networks (ANN) and A Com-bination of Feature
Selection Methods. J. Southwest Jiaotong Univ. 2019, 54. Available online:
https://ssrn.com/abstract=3619933 (accessed on 26 November 2021).
●Chen, J.-F.; Do, Q.H. Training neural networks to predict student
academic performance: A comparison of cuckoo search and gravitational
search algorithms. Int. J. Comput. Intell. Appl. 2014, 13, 1450005. [Google
Scholar] [CrossRef]
●Rashid, T.A.; Aziz, N.K. Student academic performance using artificial
intelligence. ZANCO J. Pure Appl. Sci. 2016, 28. Available online:
http://zancojournals.su.edu.krd/index.php/JPAS/article/view/544
(accessed on 26 November 2021).
●Xu, W.W.M.; Chen, X. A Performance Predictor for Honors Students
Based on Elman Neural Network in ITS 2018 Workshop Proceedings.
2018, p. 55. Available online: http://ceur-ws.org/Vol-2354/w2paper1.pdf
(accessed on 15 August 2021).
●Giannakas, F.; Troussas, C.; Voyiatzis, I.; Sgouropoulou, C. A deep
learning classification framework for early predic-tion of team-based
academic performance. Appl. Soft Comput. 2021, 106, 107355. [Google
Scholar] [CrossRef]

●Lau, E.T.; Sun, L.; Yang, Q. Modelling, prediction and classification of
student academic performance using artificial neural networks. SN Appl.
Sci. 2019, 1, 1–10. [Google Scholar] [CrossRef][Green Version]
●Rodríguez-Hernández, C.F.; Musso, M.; Kyndt, E.; Cascallar, E. Artificial
neural networks in academic performance prediction: Systematic
implementation and predictor evaluation. Comput. Educ. Artif. Intell. 2021,
2, 100018. [Google Scholar] [CrossRef]
●Sandoval-Palis, I.; Naranjo, D.; Gilar-Corbi, R.; Pozo-Rico, T. Neural
Network Model for Predicting Student Failure in the Academic Leveling
Course of Escuela Politécnica Nacional. Front. Psychol. 2020, 11, 1–7.
[Google Scholar] [CrossRef] [PubMed]
●del Río, C.; Insuasti, J.P. Predicting academic performance in traditional
environments at higher-education institutions using data mining: A
review. Ecos de la Academia 2016, 2, 185–201. [Google Scholar]
●Fan, J.; McCandliss, B.D.; Sommer, T.; Raz, A.; Posner, M.I. Testing the
eciency and independence of attentional net-works. J. Cogn. Neurosci.
2002, 14, 340–347. [Google Scholar] [CrossRef]
●Unsworth, N.; Heitz, R.P.; Schrock, J.C.; Engle, R.W. An automated
version of the operation span task. Behav. Res. Methods 2005, 37,
498–505. [Google Scholar] [CrossRef][Green Version]
●Weinstein, C.E.; Schulte, A.C.; Cascallar, E.C. The Learning and Studies
Strategies Inventory (LASSI): Initial Design and Development (Technical
Report); U.S. Army Research Institute for the Behavioral and Social
Sciences: Alexandria, VA, USA, 1982. [Google Scholar]
●Weinstein, C.E.; Palmer, D.R. LASSI User’s Manual, 3rd ed.; H & H
Publishing: Clearwater, FL, USA, 2016; Available online:
https://www.hhpublishing.com/LASSImanual.pdf (accessed on 10
August 2021).
●Weinstein, C.E.; Palmer, D.; Schulte, A.C. Learning and Study Strategies
Inventory (LASSI); H & H Publishing: Clearwater, FL, USA, 1987; Available
online:
https://asu.pure.elsevier.com/en/publications/learning-and-study-stra
tegies-inventory-lassi (accessed on 15 July 2021).
●Shahiri, A.M.; Husain, W.; Rashid, N.A. A Review on Predicting Student’s
Performance Using Data Mining Techniques. Procedia Comput. Sci. 2015,
72, 414–422. [Google Scholar] [CrossRef][Green Version]
●Huang, S.; Fang, N. Predicting student academic performance in an
engineering dynamics course: A comparison of four types of predictive
mathematical models. Comput. Educ. 2013, 61, 133–145. [Google Scholar]
[CrossRef]

●Ibrahim, Z.; Rusli, D. Predicting students’ academic performance:
Comparing artificial neural network, decision tree and linear regression.
In Proceedings of the 21st Annual SAS Malaysia Forum, Kuala Lumpur,
Malaysia, 5 September 2007. [Google Scholar]
●Asif, R.; Merceron, A.; Ali, S.A.; Haider, N.G. Analyzing undergraduate
students’ performance using educational data mining. Comput. Educ.
2017, 113, 177–194. [Google Scholar] [CrossRef]