Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1485

Personalised Learning Assistance

System for Slow Learners
Yashodha Sambrani1 (Professor) Mohan Lamani2
Computer Science & Engineering Computer Science & Engineering
SDM College of Engineering and Technology SDM College of Engineering and Technology
Dharwad, India Dharwad, India

Prasanna Kumar3 Bhavana Borah4

Computer Science & Engineering Computer Science & Engineering
SDM College of Engineering and Technology SDM College of Engineering and Technology
Dharwad, India Dharwad, India

Pragati5
Computer Science & Engineering
SDM College of Engineering and Technology
Dharwad, India

Abstract:- The Personalized Learning Assistance System irrespective of consistent classroom presence. Therefore,
(PLAS) is an innovative platform that utilizes AI/ML understanding the nuanced profiles of slow learners
algorithms to cater to the diverse needs of slow learners necessitates comprehensive assessment strategies that
in educational institutions. By analysing student learning consider not only academic progress but also attendance
behaviours, PLAS categorizes them into appropriate patterns to tailor effective interventions and support
groups and provides tailored support. Slow learners can mechanisms.
access the system through a dedicated portal, where they
can seek assistance from their peers, available teachers, National Education Policy 2020 states: “Classroom is a
or engage with e-learning videos. The platform also combination of various students, some may be from
features a portal for advanced learners, enabling them to sophisticated families, some may be from middle class and
mentor and guide their peers or access additional e- some other may be from deprived groups. In a classroom
learning resources. Furthermore, teachers can log into some students may be slow learners and some may be
the system to upload student performance data and physically or mentally challenged. Teachers understand this in
identify those who require extra support. PLAS aims to a right way and design activities to make all the students take
foster an inclusive learning environment by leveraging part in learning process. No students are excluded from the
technology to personalize the educational experience, teaching-learning process. Thus, every classroom should be
ultimately empowering slow learners to achieve academic inclusive. In the vision of the National Education Policy, it
success through a comprehensive support network. clearly says the high-quality education should be provided to
all.”
Keywords:- Personalized Assistance System, Slow Learners,
Machine Learning, Education, Adaptive Learning. Artificial intelligence (AI) and machine learning (ML)
algorithms play a pivotal role in early detection and prediction
I. INTRODUCTION of slow learners before major exams, facilitating proactive
intervention strategies to enhance academic outcomes.
Slow learners, within the educational context, are Decision tree algorithms, renowned for their interpretability,
individuals who demonstrate a prolonged or delayed analyze historical student data to identify key predictors of
progression in acquiring academic skills compared to their slow learning patterns. By leveraging features such as
peers. This term encompasses a diverse range of learners who attendance records, quiz scores, and participation levels,
may face challenges in comprehending and retaining decision trees delineate intricate decision paths to flag
information at a pace that aligns with typical developmental students at risk of underperformance. Furthermore, ensemble
expectations. Notably, slow learners may exhibit varying learning techniques like random forest harness the collective
attendance patterns, including both high attendance rates wisdom of multiple decision trees to enhance predictive
coupled with below-average academic performance, as well as accuracy. This amalgamation of diverse decision trees
low attendance rates correlating with similarly diminished mitigates overfitting and enhances generalization capabilities,
scholastic achievement. This dichotomy underscores the thereby furnishing more robust insights into students' learning
multifaceted nature of slow learning, which can manifest trajectories. Multilayer perceptron (MLP) networks, a class of

IJISRT24APR1485 www.ijisrt.com 1001

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1485

artificial neural networks, excel in modeling nonlinear harnessing the collective knowledge and expertise of both
relationships within data, thereby uncovering subtle educators and advanced learners, the personalized learning
correlations indicative of slow learning tendencies. Through assistance system not only mitigates the impact of the student-
iterative learning processes, MLPs adaptively refine their to-teacher ratio but also cultivates a dynamic and inclusive
predictive models, capturing nuanced patterns that may elude learning ecosystem conducive to academic excellence and
conventional statistical techniques. In concert, these AI/ML holistic development.
algorithms empower educators with timely alerts and
actionable insights, enabling targeted interventions and II. RELATED WORK
personalized support systems tailored to individual student
needs, ultimately fostering a conducive learning environment Govt. V.Y.T. PG Autonomous College, Durg, works on
conducive to academic success. the input required to identify slow and advanced learners such
as overall examination results, internal assessment scores, and
In today's diverse classrooms, students learn in different observations from subject teachers. This assessment is
ways, yet some face challenges keeping up with the pace of conducted systematically, with subject teachers assessing
traditional teaching methods. This project introduces a students individually for each subject in all programs.
ground-breaking approach, the Personalized Learning System Identification occurs after the declaration of preceding exam
(PLS), designed specifically for slow learners. results, considering class performance and teacher
observations. Parameters include internal assessment test
Understanding the individual needs of these students is scores (50%), academic performance in previous exams
crucial. This project focuses on utilizing cutting-edge (25%), and teacher observations (25%), recorded on a scale of
Artificial Intelligence and Machine Learning (AI/ML) 1 to 10. A report is prepared for the entire class based on these
technologies to predict and analyse learning patterns. By parameters, with students scoring below 50% identified as
applying various algorithms to live data, the PLS aims to slow learners and those scoring above 70% identified as
tailor educational experiences to suit the needs of slow advanced learners. This process aims to effectively cater to
learners. the learning needs of students by providing appropriate
support and challenges based on their academic performance
The significance of this project lies in its commitment to and observed abilities [1].
inclusivity and fairness in education. Slow learners often
struggle with conventional teaching methods, affecting their The Institute for Excellence in Higher Education
confidence and academic progress. The PLS seeks to change (IEHE), Bhopal, Madhya Pradesh, aims to provide vital
this by leveraging AI/ML technologies to predict learning facilitation and encouragement to both advanced learners to
behaviors and save data in real time. By using different excel further and slow learners to improve their performance
algorithms, it intends to personalize learning experiences, in academic and personal domains. The policy emphasizes the
offering adaptive materials and strategies. development of strategies and scientific implementations by
teaching departments to benefit all categories of learners,
This project aims to revolutionize educational including average performers. It recognizes the need to support
approaches by harnessing AI/ML techniques to understand advanced learners with demanding and challenging
and adapt to the unique learning styles of slow learners. opportunities while providing assistance and support to slow
Through predictive analysis and live data application, the PLS learners to enhance their academic and personal
endeavours to create a supportive learning environment where achievements. By addressing the diverse needs of students,
every student, regardless of their pace, can thrive and succeed. the policy seeks to create a more inclusive and effective
learning environment within the institute [2].
In the educational landscape of India, where the student-
to-teacher ratio often poses a significant challenge, leveraging Mustafa Agaoglu et al. underscores the increasing
advanced learners as part of a personalized learning assistance significance of data mining applications in addressing
system presents a promising avenue for addressing the diverse educational and administrative challenges within higher
needs of students. With limited resources and a burgeoning education. While existing research predominantly focuses on
student population, personalized attention becomes modeling student performance, this paper shifts the focus
increasingly elusive within traditional classroom settings. towards evaluating instructors' performance through the
However, by harnessing the expertise and proficiency of analysis of course evaluation questionnaires. Employing four
advanced learners, this system can effectively supplement the distinct classification techniques—decision tree algorithms,
instructional process. Advanced learners, possessing a deeper support vector machines, artificial neural networks, and
understanding of academic concepts, can serve as mentors and discriminant analysis—the study compares their efficacy in
peer tutors, offering personalized guidance and support to modeling instructor performance based on student feedback.
their peers. Through one-on-one interactions and small group Results indicate that while all classifiers demonstrate high
sessions, they can elucidate complex topics, clarify doubts, performance, the C5.0 classifier emerges as the most accurate,
and provide additional explanations tailored to individual precise, and specific. Furthermore, the analysis of variable
learning styles. Moreover, their role extends beyond mere importance reveals the presence of irrelevant questionnaire
academic assistance; they also serve as role models, inspiring items, emphasizing the pivotal role of student interest in
their peers through their academic achievements and fostering determining instructor success as perceived by students. These
a culture of collaborative learning and mutual support. By findings underscore the efficacy and relevance of data mining

IJISRT24APR1485 www.ijisrt.com 1002

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1485

models in enhancing course evaluation practices and predicting student attendance and comparing results with slow
advancing insights into higher education dynamics, offering learner students' performance [6].
valuable insights for refining measurement instruments [3].
Sameh S. Alfere et al. delve into the burgeoning field of
Ijaz Khan et al. address a significant challenge faced by educational research utilizing Data Mining techniques, which
instructors: effectively monitoring students' academic has seen a rapid increase. Focusing on uncovering hidden
progress in courses. They propose leveraging machine patterns and knowledge within student performance data,
learning algorithms to identify students with unsatisfactory their study centers on predicting student performance through
progress early on, allowing instructors to provide additional classification methods, specifically modified KNN classifiers
support promptly. Despite the abundance of student data like Cosine KNN, Cubic KNN, and Weighted KNN. Utilizing
collected by modern educational institutions, there is a need a dataset from 11th-grade students in the scientific branch of
for innovative methods to utilize this data to enhance Gaza Strip secondary schools, comprising 13 parameters
educational quality and institutional prestige. The research including subject marks, average, and grade, the classifiers
evaluates the effectiveness of various machine learning aim to forecast the grade a student will achieve based on their
algorithms in predicting students at risk of poor performance marks in two subjects. This predictive analysis holds promise
and transforms the prediction model into an easily for aiding school principals in early identification of students
interpretable format for instructors. Decision tree algorithms at risk of low academic achievement, enabling timely
are found to be most successful and are refined for practical interventionand support strategies [7].
use. The outcome of the research includes a set of proactive
measures for monitoring students' performance from course Parneet Kaur et al. focus on the application of predictive
inception and providing targeted support to struggling data mining models using classification algorithms to identify
students [4]. slow learners among students and monitor curriculum changes
in educational databases. Utilizing a real-world dataset from a
K. Hemachandran et al. delve into the transformative high school and WEKA, an Open Source Tool, they filter
potential of artificial intelligence (AI) in higher education, relevant variables and test various classification algorithms
highlighting its underutilization among professors despite its including Multilayer Perception, Naïve Bayes, SMO, J48,
revolutionary impact on human lives. They emphasize the and REPTree. The paper compares statistics generated by
urgent need for implementing information bridge technology these algorithms to predict accuracy and determine the best-
to enhance communication in classrooms and predict the performing classification algorithm. Additionally, a
future of higher education with AI. Through this research knowledge flow model among the classifiers is illustrated,
article, the authors address the current challenges faced by emphasizing the significance of prediction and classification-
subject faculties and students, including changing government based data mining algorithms in education while suggesting
regulations, while exploring various arguments and hurdles future research directions [8].
surrounding AI implementation in education. They present a
use case model leveraging student assessment data and Ismail Celik et al. provide an overview of research
generative adversarial networks (GAN) to bridge the gap examining teachers' utilization of artificial intelligence (AI)
between human lecturers and machines, achieving a applications and machine learning methods for analyzing
maximum accuracy of 58% using various machine learning teacher data. Their analysis reveals numerous opportunities
algorithms. The study also considers the psychological impact AI presents for teachers, including improved planning by
of AI on faculty and students, aiming to ensure a balanced defining students' needs and facilitating teacher familiarity
integration of technology while preserving human emotions with them, enhanced implementation through immediate
and interactions in educational settings [5]. feedback and intervention, and streamlined assessment via
automated essay scoring. Moreover, the study highlights
S. Deepa et al. explore the application of data mining teachers' diverse roles in AI technology development, ranging
(DM) in education, focusing on the emerging from serving as models for AI algorithm training to
interdisciplinary field known as educational data mining participating in AI system accuracy checks. However, the
(EDM). Their research aims to enhance student performance authors also identify several challenges in implementing AI
by improving the quality of educational processes through in teaching practice, offering valuable insights for the field's
new strategies and plans. They address the challenges of development [9].
exploring unique data types from educational environments,
particularly focusing on slow learner students in school Sangeeta. K et al. emphasize that increasing the pass
education. Factors such as economy, environment, social percentage of students is a primary goal for educational
media, psychological issues, childhood marriage, irregular institutions, acknowledging that a student's performance is
attendance, and eve-teasing of girls are identified as influenced by various factors, including their learning ability.
influencing factors. The goal is to gain insights into how They highlight the distinction between slow learners, who
students learn and identify settings for improvement, utilizing require detailed explanations and resources to grasp concepts,
data mining techniques such as the J48 algorithm to predict and fast learners, who assimilate information more quickly.
students at risk of not attending school. Through attribute Given the competitive demands of the modern world,
selection and data rebalancing using cost-sensitive classifying students based on their learning abilities can aid in
classification, the study aims to improve accuracy in identifying slow learners who may benefit from additional
support and training to improve their performance. To address

IJISRT24APR1485 www.ijisrt.com 1003

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1485

this, the paper utilizes real-time student data from the classification algorithms, the study aims to categorize students
computer science engineering department at Aditya Institute into different learning groups and develop a predictive model
Of Technology and Management, Tekkali, Srikakulam to better understand individual differences in knowledge
district, to conduct experiments on the influence of cognitive levels, learning preferences, and cognitive abilities [10].
attributes on academic performance. By employing

Table 1 Literature Survey

S. l Title Method Accuracy
1. Predicting Decision Tree, 90%
Instructor Performanc e Using SVM
Data Mining Techniques in
Higher Education
2. An artifcial KNN, WEKA(Decision Decision Tree- 93%
intelligence approach to monitor tree), ANN, Naive Bayes
student performanc e and devise
preventive measures
3. Artificial SVM, KNN, Logistic Multilayer Perception - 87.44%
Intelligence : A Regression (LR), Linear Discriminant
Universal Virtual Tool to Analysis (LDA), K-Nearest Neighbors
Augment Tutoring in Higher (KNN), Regression Trees (CART),
Education Naive Bayes (NB), Support Vector
Machines (SVM), Random Forest (RF)
4. Identificatio Naïve Bayes, J48 - 97%
n of slow learners with J48 Algorithm using SQL tool
categorizati on data mining
techniques
5. Prediction Fine, Medium, Coarse, Weighted KNN
of Student's Performanc e Using Cosine KNN, Cubic KNN Classifier, Classifier-94.1%
Modified KNN Weighted KNN Classifier
Classifiers
6. Classificati Decision Tree (DT), Multi-Layer Perception-75%,
on and prediction based data Random Forest (RF), Neural Network Naïve Bayes-65.13%,
mining algorithms to predict slow (NN) and Support Vector Machine SMO-68.42%,
learners in education sector (SVM) J48-69.73%, REPTree-67.76%
7. Classification and Prediction of CHAID, SVM, Multilayer Perception, CHAI-59.4% , SVM-80% ,
Slow Learners Using Machine Random Forest, REPTree, J48, SMO, Multilayer perception-75%, RandomForest-
Learning Algorithms Naive Bayes 75%, REPTree-54.4%,
J48-65.8%, SMO-58.5%, Naive Bayes-55.8%

III. CLASSIFICATION  A decision tree is a simple yet powerful machine learning

algorithm used for classification and regression tasks. It
 Three Algorithm’s we chose after Thorough Literature organizes data into a tree-like structure where each branch
Survey: represents a decision based on features, leading to
predicted outcomes at the tree's leaf nodes. It's easy to
 Random Forest is a popular and powerful ensemble interpret, but can over fit with complex data—techniques
learning method used in machine learning. It operates by like pruning help prevent this. Overall, decision trees are
constructing multiple decision trees during training and effective tools for making decisions based on data
outputs the mode of the classes (classification) or average patterns.
prediction (regression) of the individual trees. It's an  A Multilayer Perceptron (MLP) is a type of artificial
ensemble technique that combines the predictions from neural network with layers of interconnected nodes
several individual decision trees to generate a more (neurons). It's structured in an input layer, hidden layers,
accurate and robust final prediction. and an output layer. MLPs learn patterns in data by
adjusting connections between neurons through a process
Ensemble Learning refers to the technique of called backpropagation. They're widely used in various
combining multiple individual models (learners) to build a fields for tasks like pattern recognition, classification, and
stronger, more accurate model. The basic idea is that a group prediction due to their ability to learn complex
of weak learners can come together to form a strong learner. relationships in data.
Ensemble methods often outperform single models because
they reduce variance, increase accuracy, and mitigate
overfitting by aggregating the predictions of multiple models.

IJISRT24APR1485 www.ijisrt.com 1004

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1485

IV. PROPOSED MODEL

Fig 1 Detailed Design of Our Model

 EDM Model refinement, thereby fostering continuous improvement and

At the heart of the learning system lies the EDM model, innovation within the educational ecosystem.
a sophisticated framework responsible for not only modeling
the intricate learning process but also for dynamically  Classifying and Categorizing Students
adapting to the unique needs and preferences of individual Through the classifying and categorizing students
students. Employing a diverse array of artificial intelligence module, learners are systematically grouped based on their
(AI) and machine learning (ML) techniques such as ensemble distinct learning needs and aptitudes. This segmentation
learning and deep learning, the EDM model serves as the enables the EDM model to craft personalized learning
cornerstone of personalized learning experiences. pathways tailored to the unique requirements of each student,
fostering a supportive and inclusive learning environment.
 Learner Modelling Database
The comprehensive learner modelling database serves  Advanced Learners
as a repository of invaluable information pertaining to each Advanced learners are afforded challenging tasks and
student. From prior knowledge and academic details to activities that stimulate their intellectual curiosity and foster
performance metrics across various subjects, this database their growth. Encouraged to pursue independent projects and
encapsulates the essence of each learner's educational share their expertise with peers, advanced learners serve as
journey. Leveraging this rich dataset, the EDM model catalysts for knowledge dissemination and peer-to-peer
meticulously tailors learning experiences to suit the specific learning within the classroom.
requirements of every student, ensuring optimal engagement
and academic growth.  Consistent Learners
Consistent learners benefit from activities and tasks
 MIS Data designed to sustain and reinforce their academic progress.
The Management Information System (MIS) data Through opportunities for collaborative learning and
provides invaluable insights into the overall performance and engagement with peers and educators, consistent learners are
efficacy of the learning system. By analyzing this data, empowered to build upon their existing knowledge and skills
educators can pinpoint areas for enhancement and in a supportive and nurturing environment.

IJISRT24APR1485 www.ijisrt.com 1005

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1485

 Slow Learners The weightage division can be used to identify students

Recognizing the diverse needs of slow learners, the who need extra support, develop personalized learning plans,
system provides additional support and guidance tailored to and track student progress over time.
their individual learning pace. Emphasizing personalized
learning experiences and opportunities for remediation, slow Table 2 Weightage Division as a Parameter
learners are equipped with the tools and resources necessary
to overcome challenges and thrive academically.

 Feedback
A robust feedback mechanism, facilitated by the EDM
model, educators, and peers, enables students to receive
constructive feedback on their performance. By leveraging
this feedback, students can identify areas of strength and
areas for improvement, empowering them to develop
effective learning strategies and achieve their full potential.
The weightage division shown in above table allocates
 Flexibility and Scalability different weights to different factors that influence a student's
Designed with flexibility and scalability in mind, the e- performance, such as academics, attendance, and behaviour.
learning system is poised to adapt and evolve over time. With This helps identify students who need extra support and tailor
the ability to integrate new data and cutting-edge AI and ML interventions to their specific needs. It also allows for
techniques, the system continually refines its performance, tracking progress over time to evaluate the effectiveness of
ensuring optimal outcomes for students across diverse the interventions.
learning environments and scales.
The Slow Learner Assistance Portal is designed for both
educators and students offers a sophisticated interface
optimized for educational management. Upon authentication,
students gain access to a curated list of courses corresponding
to their enrollment. Within this interface, students navigate to
their chosen subject, initiating a streamlined educational
journey.

Students who are slow learners in their selected subject

are seamlessly directed to an interface offering tailored
support options. These include peer-to-peer engagement for
collaborative learning opportunities, scheduled meetings with
educators for targeted remediation sessions, and access to a
repository of e-learning materials designed to augment
comprehension and retention.

Students who are advanced learners in their selected

subject also have access to e-learning resources, ensuring a
comprehensive educational experience that caters to diverse
proficiency levels.

For educators, the portal streamlines the assessment and

tracking process. Through an intuitive interface, educators
can directly upload marks of the whole class onto the
Fig 2 Student Classifying Categories platform. Upon submission, educators have the option to
generate a downloadable report. This report categorizes
 The Figure 2. Shows Three Levels of Students: Advanced student performance based on predefined metrics, facilitating
Learner, Consistent Learner, and Slow Learner. informed decision-making and personalized intervention
strategies.
 Students are classified based on their SEE/2 + CIE score
and attendance. V. RESULTS
 Slow Learners are students with a SEE/2 + CIE score
less than 60. In conducting a comparative study of three machine
 At-Risk Students Are Slow Learners with attendance less learning algorithms—random forest, decision tree, and
than 40%. multilayer perceptron (MLP)—across varying epochs and
 Underachievers are Slow Learners with attendance 70% train-test divisions, our aim was to ascertain the optimal
or higher. configuration for predicting student performance and

IJISRT24APR1485 www.ijisrt.com 1006

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1485

identifying slow learners within an educational context. We 20. Each configuration underwent rigorous testing through
systematically evaluated each algorithm's performance across three trial runs to ensure robustness and reliability of the
different epoch counts, ranging from 20 to 100, while also results, as shown in the below mentioned table.
varying the train-test division ratios at 60-40, 70-30, and 80-

Table 3 Comparative Study between the Models

Across all epochs and train-test division ratios, random epochs exhibited notable enhancements in predictive
forest consistently demonstrated superior predictive accuracy performance. This trend underscores the importance of
compared to decision tree and MLP. Notably, the random sufficient epoch counts in facilitating convergence and
forest algorithm yielded the highest accuracy rates when refining the model's predictive capacity over successive
trained and tested with an 80-20 split, indicating its efficacy iterations.
in handling larger training datasets while maintaining
generalization capabilities. Conversely, decision tree and MLP algorithms, while
demonstrating commendable performance, consistently fell
Specifically, upon conducting three trials for random short of random forest's accuracy levels across all epoch
forest with 80-20 train-test division, we observed a counts and train-test division ratios. Despite variations in
discernible trend towards improved accuracy with increasing epoch counts and train-test splits, decision tree and MLP
epoch counts. While the initial trial at 20 epochs yielded algorithms exhibited relatively stable performance, albeit
respectable accuracy, the subsequent trials at 50 and 100 with lower accuracy rates compared to random forest.

Fig 3 An Example of Student Classified into Categories by Our Model

IJISRT24APR1485 www.ijisrt.com 1007

Volume 9, Issue 4, April – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24APR1485

VI. CONCLUSION [11]. Abdulazeez Yusuf et al.. “Prediction of Students’

Academic Performance Using Educational
The challenges faced by educational institutions in Datamining Technique: Literature Review”,
addressing the needs of slow learners underscore the dissertation_paper-libre.
necessity for a personalized assistance system. Such a [12]. S. E. Sorour, T. Mine, K. Goda, and S. Hirokawa, J.
system, capable of dynamically adapting to the specific Inf. Process., 23, 2, 192–201 (2015).
requirements of slow learners, holds immense potential to [13]. A. Sharabiani, F. Karim, A. Sharabiani, M. Atanasov,
enhance their learning outcomes. The objectives outlined, and H. Darabi, IEEE Glob. Eng. Educ. Conf., 832–
including the identification and classification of students, 8372 (2014).
comparative analysis of machine learning models, provision [14]. H. P. Tuckman, J. Econ. Educ., 7, 1, 34–39 (1975).
of personalized learning plans, and continuous progress [15]. V. C. Smith, A. Lange, and D. R. Huston, J.
monitoring, form a comprehensive approach towards Asynchronous Learn. Networks, 16, 3, 51–61 (2012).
addressing this critical need. Through a thorough literature [16]. Y. Itoh, H. Itoh, and F. Kenji, IEEE 4th Global
survey and comparative analysis, the groundwork has been Conference on Consumer Electronics (GCCE), 288–
laid for the development of an effective personalized 291 (2015).
assistance system. System development, focusing on building [17]. A. Abdul Aziz, N. H. Ismail, and F. Ahmad,
a secure database and user-friendly interface, and Proceeding of the International Conference on
implementing predictive algorithms as the backend, is crucial Artificial Intelligence and Computer Science(AICS),
for the successful execution of personalized learning and 15–16 (2014).
remedial suggestions. Furthermore, the establishment of [18]. A. Mohamed Shahiri, W. Husain, and N. Abdul
assessment metrics and continuous feedback loops ensures Rashid, Procedia Comput. Sci., 72, 414–422 (2015).
the ongoing refinement and improvement of the system,
ultimately contributing to a more inclusive and effective
educational environment for slow learners.

REFERENCES

[1]. Govt. V.Y.T. PG Autonomous College, Durg,

“Guidelines for Identification and Activities for Slow
and Advanced Learners of Institution”, Internal
QaulityAssurance Cell (IQAC).
[2]. Government College Bichhua, Chhindwara, M.P,
“Policy on Advanced Learner & Slow Learner”,
HigherEducation.mp.gov.in, May 2017.
[3]. Mustafa Agaoglu, “Predicting Instructor Performance
Using Data Mining Techniques in Higher Education”,
IEEE Access, June 2016.
[4]. Ijaz Khan et al., “An artifcial intelligence approach to
monitor student performance and devise preventive
measures”, Springer Open, 2021.
[5]. K. Hemachandran et al., “Artificial Intelligence: A
Universal Virtual Tool to Augment Tutoring in Higher
Education”, Hindawi, 2022.
[6]. S. Deepa et al., “Identification of Slow Learners with
Categorization Data Mining Techniques”, International
Journal of Pure and Applied Mathematics.
[7]. Sameh S. Alfere et al., “Prediction of Student's
Performance Using Modified KNN Classifiers”,
International Conference On Engineering & Future
Technology, 2018.
[8]. Parneet Kaur et al., “Classification and Prediction
Based Data Mining Algorithms to Predict Slow
Learners in Education Sector”, Elsevier, Vol 57, 2015.
[9]. Ismail Celik et al., “The Promises and Challenges of
Artifcial Intelligence for Teachers: a Systematic
Review of Research”, TechTrends Springer, 2022.
[10]. Sangeeta. K et al., “Classification and Prediction of
Slow Learners Using Machine Learning Algorithms”,
International Journal of Computer Trends and
Technology, Vol- 68, 2020.

IJISRT24APR1485 www.ijisrt.com 1008