International Journal of Modern Education and Computer Science @ijmecs
Статьи журнала - International Journal of Modern Education and Computer Science
Все статьи: 1112
Статья обзорная
This study presents the development and evaluation of a local agent-based Retrieval-Augmented Generation (Agentic RAG) system designed for the intelligent analysis of GitHub repositories in computer science education and IT practice. The novelty of this work lies not in inventing a new RAG algorithm, but in orchestrating multiple existing components (LangChain, Redis, SentenceTransformer, and LLMs) into a multi-stage agent pipeline with integrated relevance evaluation, specifically adapted to offline repository mining. The proposed pipeline consists of four sequential stages: (1) query reformulation by a dedicated LLM agent, (2) semantic retrieval using SentenceTransformer embeddings stored in Redis, (3) response generation by a second LLM, and (4) relevance scoring through a verification agent with retry logic. Relevance is assessed via cosine similarity and LLM-based scoring, allowing iterative refinement of answers. Experimental testing compared the system against two baselines: keyword search and a non-agentic single-stage RAG pipeline. Results showed an average MRR@10 of 0.72, compared to 0.48 for keyword search and 0.61 for non-agentic RAG, representing a 33% relative improvement in retrieval quality. Human evaluators (n=15, computer science students) rated generated explanations on a 5-point Likert scale; the proposed system achieved an average 4.3/5 for clarity and correctness, compared to 3.6/5 for the baseline. Precision@5 for code retrieval improved from 0.54 (keyword) and 0.67 (non-agentic RAG) to 0.76 in the proposed system. Average query latency in the local environment was 3.8 seconds, indicating acceptable performance for educational and small-team IT use cases. The system demonstrates high autonomy by operating fully on-premises with only optional API access to LLMs, ensuring privacy and independence from cloud providers. Ease of use was measured through a System Usability Scale (SUS) questionnaire, yielding a score of 78/100, reflecting positive user perception of the Streamlit interface and minimal setup requirements. Nevertheless, several limitations were observed: the high computational cost of running embeddings and LLMs locally, potential hallucinations in generated explanations (particularly for complex or unfamiliar code), and the inability of vector search to fully capture code syntax and control flow structures. Furthermore, while the Analytic Hierarchy Process (AHP) was applied to select the system architecture, future work should complement this with benchmark-driven evaluations for greater objectivity. The contribution of this study is threefold: (1) introducing a multi-agent orchestration logic tailored to educational code repositories; (2) empirically demonstrating measurable gains in retrieval quality and explanation usefulness over baselines; and (3) highlighting both opportunities and limitations of deploying autonomous RAG systems locally. The proposed technology can benefit IT companies seeking secure in-house tools for repository analysis, universities aiming to integrate intelligent assistants into programming courses, and research institutions requiring reproducible, privacy-preserving environments for code exploration.
Бесплатно
Log-File Analysis to Identify Internet-addiction in Children
Статья научная
The problem of the Internet addiction (IA) arose after the rise of the Internet. Some of the Internet users include children and teenagers and they are active in a virtual environment. Most minor users are not well aware of the dangers posed by information abundance. One of these dangers is the IA. Excessive use of the Internet is addictive, and some users experience a high risk of addiction. IA can negatively affect the children's health, psychology, socialization and other activities. There is a great need to the development of forecasting programs and various technological approaches for the identification of IA among Internet users, especially children and adolescents. This article uses machine-learning techniques to detect IA. Activities of children in the Internet environment is analyzed. The log-files of children and their IA problem are explored. To determine the degree of IA among children and adolescents an experiment is conducted on public dataset. The effectiveness of the methods is analyzed by various evaluation metrics and promising results are obtained.The results show better performance of Weighted SVM, compared to BernoulliNB, Logistic Regression, MLPClassifier, SVM classifiers. Acquired results of the research provide kids information security. To evaluate a kids IA helps to identify their psychological conditions, and it creates a better situation for parents, teachers, and other related people to communicate with children and teenagers better way.
Бесплатно
Lossy Compression of Color Images using Lifting Scheme and Prediction Errors
Статья научная
This paper presents an effective compression technique for lossy compression of color images. After reducing the correlation among R, G and B planes using YCoCg-R transform, the Integer Wavelet Transform (IWT) is applied on each of the transformed planes independently up to a desired level. IWT decomposes the input image into an approximation and several detail subbands. Approximation subband is compressed losslessly using prediction errors and Huffman coding, while each of the detail subbands are compressed independently using an effective quantization and Huffman coding. To show the effectiveness of proposed scheme, it is compared with several existing schemes and a state of art for image compression JPEG2000 and it is observed that the proposed scheme outperforms over the existing techniques and JPEG2000 with less degradation in the quality of reconstructed images while achieving high compression performance.
Бесплатно
Статья научная
Microcontrollers and FPGAs both are widely used in digital system design. Microcontroller-based instruments are becoming increasingly widespread. This paper presents design and implementation of a new low-cost and minimum embedded board based on ATMEGA32L AVR microcontroller and Spartan-3 (XCS400-4PQG208C) FPGA in two layers with mount elements on top and button of board. Using of AVR microcontroller in proposed board it adds many features include Analog to Digital Converter (ADC), Digital to Analog Converter (DAC), 32 Kbytes flash memory, 2 Kbytes SRAM, 1024 bytes EEPROM memory. The design goal was to implement as many as possible low-cost and minimum size of the board, also to receive and process input signals in a short time period as real time. The board features are; mount elements in two side of the board for minimization of proposed board and also place decoupling capacitors (by pass) for the FPGA in bottom layer of board strictly below this IC because they should be placed as close as possible to the power supply pins FPGA, use GND polygon layer in total top layer and microcomputer ground for FPGA in bottom layer, use two RS-232 serial port, one VGA connector, PS/2 serial port, and SPI serial port on FPGA, use MT48LC16M16A SDRAM-256MB(4*4MB*16), and XCF02S configuration PROM. Size of the proposed embedded board is 10cm*15cm thus this board was optimized of aspect cost, performance, power, weight, and size.
Бесплатно
MLP based Reusability Assessment Automation Model for Java based Software Systems
Статья научная
Reuse refers to a common principle of using existing resources repeatedly, that is pervasively applicable everywhere. In software engineering reuse refers to the development of software systems using already available artifacts or assets partially or completely, with or without modifications. Software reuse not only promises significant improvements in productivity and quality but also provides for the development of more reliable, cost effective, dependable and less buggy (considering that prior use and testing have removed errors) software with reduced time and effort. In this paper we present an efficient and reliable automation model for reusability evaluation of procedure based object oriented software for predicting the reusability levels of the components as low, medium or high. The presented model follows a reusability metric framework that targets the requisite reusability attributes including maintainability (using the Maintainability Index) for functional analysis of the components. Further Multilayer perceptron (using back propagation) based neural network is applied for the establishment of significant relationships among these attributes for reusability prediction. The proposed approach provides support for reusability evaluation at functional level rather than at structural level. The automation support for this approach is provided in the form of a tool named JRA2M2 (Java based Reusability Assessment Automation Model using Multilayer Perceptron (MLP)), implemented in Java. The performance of JRA2M2 is recorded using parameters like accuracy, classification error, precision and recall. The results generated using JRA2M2 indicate that the proposed automation tool can be effectively used as a reliable and efficient solution for automated evaluation of reusability.
Бесплатно
MLRTS: multi-level real-time scheduling algorithm for load balancing in fog computing environment
Статья научная
Cloud computing is an innovative technology which is based on the internet to preserve large applications. It is warehoused as a shared data over one platform. In addition, it offers better services to clients who belong to different organizations. In spite of the maximum utilization of computational resources provided by the cloud computing with lower cost, it suffers from specific restrictions. These restrictions are encountered through the load balancing of data in the cloud data centers. These restrictions are represented in the less bandwidth utilization, resource limitations, fault tolerance and security etc. In order to overcome these limitations, new computing model called Fog Computing is presented. It aims to offer the required service of the sensitive data to end users without delaying. The function of the fog computing is similar to the cloud computing with two preferred advantages. The first one is that it is placed more near to the end users to introduce its service in less time. Secondly, it is more valuable for streaming the real time applications, sensor networks, IOT which need high speed and reliable internet connection. In this paper, a novel load balancing algorithm has been proposed over a novel architectural model in the Fog Computing environment. The proposed model aims to serve the real-time tasks within their deadline. In addition, it serves the different soft tasks without starving. The soft tasks are classified according to the execution time and the priority levels. In addition, they are served according to their waiting time and priority-level. Furthermore, the proposed algorithm is employed to maximize the throughput, the resources and the network utilization and preserving the data consistency with less complexity to accomplish the end users demand.
Бесплатно
Machine Learning Algorithms for Quantifying the Role of Prerequisites in University Success
Статья научная
The use of machine learning algorithms for higher education performance assessment is an emerging area of research and several works have focused on student performance and related problems. The preliminary goal of this work is to determine and quantify the role of prerequisites in academic success by using machine learning algorithms with the Weka environment. The main objective is the development of a tool based on machine learning algorithms for the prediction of future results for a training program based solely on the previous academic profiles of the students. The interest is to link whether success in previous courses is associated with success in subsequent target courses. This will help to improve the planning of course sequences in a training program on the one hand and the overall academic students’ success on the other. The proposed methodology is applied for the analysis of the role of the prerequisites influencing courses success of a training course in Mathematical and Computer Sciences in a Moroccan university. For this purpose, we use several classification algorithms such as Random Forest, J48, and Multilayer Perceptron. Preliminary results show that the correlation between the prerequisite reliability rates of the courses studied and the accuracy with which the learning algorithms predict the success outcomes of these courses is confirmed. Also, these results show that the best accuracy and the best Receiver Operator Characteristic ROC area are obtained by using Random Forest algorithm and have reached 86% for the accuracy and 75.6% for the ROC area.
Бесплатно
Machine Learning Cross Layer Technique to Detect Sink Hole Attacks in MANET
Статья научная
Adhoc networks uses mobile nodes to communicate among itself in which it does not have any fixed infrastructure like access point or base station. Due to dynamic network topology MANET security is a challenging task. Most of the routing protocols in MANET assumes a cooperative environment for communication. But, in the presence of malicious nodes, providing security to MANET is critical issue. Due to the increasing applications of MANET building an effective intrusion detection system are essential. This paper addresses using an intelligent approach for intrusion detection in MANET using cross layer technique. We show an paradigm of SVMs, FDAs and AIS approaches for intrusion detection in terms of classification accuracy.
Бесплатно
Machine Learning Elman Technique for Predicting Shelf Life of Burfi
Статья научная
Elman artificial neural network single and multilayer computerized models were developed for predicting the shelf life of burfi stored at 30ºC. The experimental data of the product relating to moisture, titratable acidity, free fatty acids, tyrosine, and peroxide value were taken as input variables, and overall acceptability score as output variable for developing the models. Bayesian regularization algorithm was applied as training algorithm for neural network. Transfer function for hidden layers was tangent sigmoid; while for output layer it was pure linear function. Elman model with a combination of 5→10→1 and 5→7→7→1 performed exceedingly well for predicting the shelf life of burfi.
Бесплатно
Статья научная
The Question-Answering (QA) approach represents one of the most significant Natural Language Processing (NLP) tasks that depends on language input. In terms of morphology & adhesive structure, Malayalam is a resource-constrained indigenous language of India. These linguistic features make QA in Malayalam particularly difficult. This study uses a subset of 5 tasks from the Facebook bAbI dataset to present a subset of five assignments from the Facebook bAbI dataset; this study presents a Malayalam Question Answering Solution that utilizes a Deep Learning (DL) hybrid framework combining CNN and Bi-LSTM Methods. We believe this is the initial time a hybrid-based deep learning framework has been used for the Malayalam question-answering technology. In the first iteration of the method, high-level semantic characteristics are extracted utilizing a Convolutional Neural Network. The Bi-LSTM tier then extracts the contextual feature representation of the text using the feature extraction result. Finally, use the softmax activation function to predict correct answers for corresponding questions. The proposed model is both functional and systemized in terms of classification accuracy, precision, recall, and F1 scores. The simulation results show that the proposed hybrid CNN and Bi-LSTM model outperform the existing models in terms of classification with more than 91 % accuracy for all five tasks.
Бесплатно
Статья научная
Significant research into the logarithmic analysis of complex networks yields solution to help minimize virus spread and propagation over networks. This task of virus propagation is been a recurring subject, and design of complex models will yield modeling solutions used in a number of events not limited to and include propagation, dataflow, network immunization, resource management, service distribution, adoption of viral marketing etc. Stochastic models are successfully used to predict the virus propagation processes and its effects on networks. The study employs SI-models for independent cascade and the dynamic models with Enron dataset (of e-mail addresses) and presents comparative result using varied machine models. Study samples 25,000 emails of Enron dataset with Entropy and Information Gain computed to address issues of blocking targeting and extent of virus spread on graphs. Study addressed the problem of the expected spread immunization and the expected epidemic spread minimization; but not the epidemic threshold (for space constraint).
Бесплатно
Mathematical Framework for A Novel Database Replication Algorithm
Статья научная
In this paper, the detailed overview of the database replication is presented. Thereafter, PDDRA (Pre-fetching based dynamic data replication algorithm) algorithm as recently published is detailed. In this algorithm, further, modifications are suggested to minimize the delay in data replication. Finally a mathematical framework is presented to evaluate mean waiting time before a data can be replicated on the requested site.
Бесплатно
Mathematical Model for Adaptive Technology in E-learning Systems
Статья научная
The emergence of a large number of e-learning platforms and courses does not solve the problem of improving the quality of education. This is primarily due to insufficient implementation or lack of mechanisms for adaptation to the individual parameters of the student. The level of adaptation in modern e-learning systems to the individual characteristics of the student makes the organization of human-computer interaction relevant. As the solution of the problem, a mathematical model of the organization of human-computer interaction was proposed in this work. It is based on the principle of two-level adaptation that determines the choice of the most comfortable module for studying at the first level. The formation of an individual learning path is performed at the second level. The problem of choosing an e-module is solved using a fuzzy logic. The problem of forming a learning path is reduced to the problem of linear programming. The input data are the characteristics of the quality of student activity in the education system. Based on the proposed model the computer technology to support student activities in modular e-learning systems is developed. This technology allows increasing the level of student’s cognitive comfort and optimizing the learning time. The most important benefit of the proposed approach is to increase the average score and increase student satisfaction with learning.
Бесплатно
Mathematical Modeling and Analysis of Network Service Failure in DataCentre
Статья научная
Malik UsmanDilawar, FaizaAyub Syed
Бесплатно
Mathematics Is Science: A Topic Revisited in Context of FCS of India
Статья научная
Mathematics is universally accepted as mother of all science. Despite that, Department of Personnel and Training (DOPT) has recently issued a circular mentioning that a person having master degree in mathematics cannot be considered for the post of scientists. The open question of 'Is mathematics a science?' is revisited in this paper under the new perspective to explore scientific practices that sans mathematics arrived knocking, challenging basic understanding of precision and practical sense that makes science. Considering the fact that in India, most crucial policy decisions at a higher level of abstraction in every conceivable arena of our national life are taken by either GOM (Group of Ministers) or GOS (Group of Secretaries), apprehension raises a basic query 'Who decides?' Some decision causes much unexpected consequence, which is noticed when it takes its toll and becomes virtually irreversible. This recent decision of Flexible Complementing Scheme (FCS), wherein mathematics is not considered as science, has potential to damage the very scientific culture and practices in India. This paper is an attempt to place mathematics in its right perspective and to highlight the damage that this decision might do. The paper also suggests ways to control the damage.
Бесплатно
Matrix Based Energy Efficient Scheduling With S-MAC Protocol in Wireless Sensor Network
Статья научная
Communication is the main motive in any Networks whether it is Wireless Sensor Network, Ad-Hoc networks, Mobile Networks, Wired Networks, Local Area Network, Metropolitan Area Network, Wireless Area Network etc, hence it must be energy efficient. The main parameters for energy efficient communication are maximizing network lifetime, saving energy at the different nodes, sending the packets in minimum time delay, higher throughput etc. This paper focuses mainly on the energy efficient communication with the help of Adjacency Matrix in the Wireless Sensor Networks. The energy efficient scheduling can be done by putting the idle node in to sleep node so energy at the idle node can be saved. The proposed model in this paper first forms the adjacency matrix and broadcasts the information about the total number of existing nodes with depths to the other nodes in the same cluster from controller node. When every node receives the node information about the other nodes for same cluster they communicate based on the shortest depths and schedules the idle node in to sleep mode for a specific time threshold so energy at the idle nodes can be saved.
Бесплатно
Mean-Field Theory in Hopfield Neural Network for Doing 2 Satisfiability Logic Programming
Статья научная
The artificial neural network system's dynamical behaviors are greatly dependent on the construction of the network. Artificial Neural Network's outputs suffered from a shortage of interpretability and variation lead to severely limited the practical usability of artificial neural networks for doing the logical program. The goal for implementing a logical program in Hopfield neural network rotates rounding minimizing the energy function of the network to reaching the best global solution which ordinarily fetches local minimum solution also. Nevertheless, this problem can be overcome by utilizing the hyperbolic tangent activation function and the Boltzmann Machine in the Hopfield neural network. The foremost purpose of this article is to explore the solution quality obtained from the Hopfield neural network to solve 2 Satisfiability logic (2SAT) by using the Mean-Field Theory algorithm. We want for replacing the real unstable prompt local field for the separate neurons into the network by its average local field utility. By using the solution to the deterministic Mean-Field Theory (MFT) equation, the system will derive the training algorithms in which time-consuming stochastic measures of collections are rearranged. By evaluating the outputs of global minima ratio (zM), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) with computer processing unit (CPU) time as benchmarks, we find that the MFT theory successfully captures the best global solutions by relaxation effects energy function.
Бесплатно
Measures for the Ontological Relations in Enterprise
Статья научная
In order to improve a system performance, it is significant to estimate the exchange rate of relationships between components of the system, in particular when the considered system is production or service companies. Indeed, bad and inappropriate relationships can generate dysfunctions, slowdowns or, more generally, loss of performance in enterprise leading to a decline in growth and competitiveness. Because of the heterogeneity of information and data, it is necessary to modeling relationships and ontologies are currently among the most evoked models in knowledge engineering. The aim is to define structured vocabularies, bringing together useful concepts of a domain and their relationships thus serving to organize, exchange information in an unambiguous way. Ontologies are widely applied to ensure semantic interoperability describing the enterprise structure and the exchange rate of existing relationships can be valued through their degree of effectiveness. This paper presents measures for the ontological relations in the enterprise. Our approach aims first to extract the set of relationships from an ontology previously created, then classify these relations, according to two types giving a weighting to calculate their degree of effectiveness. The implementation process is proposed on the local enterprise of steel wire drawing processing, giving degree of effectiveness for existing relationships. A sensitivity analysis is done to compare and interpret the different results.
Бесплатно
Measuring of Software Maintainability Using Adaptive Fuzzy Neural Network
Статья научная
Software maintenance mainly refers to the process of correcting software after delivery. Maintenance process is usually a high percentage of Organizational effort to the whole process of software programs. As a result, the effectiveness of the entire production process and customer satisfaction is dependent on the effectiveness of maintenance activities. Because many factors including type of service, type of product and human factors is dependent on the maintenance process, And the imprecise nature of qualitative factors and sub-criteria leading software maintenance, accurate assessment can be maintained in order to measure the effectiveness of programs seem highly desirable. In this paper, using adaptive fuzzy neural network to provide a method for evaluating the capability of software maintenance conducted after the tests, the root mean square error of the proposed method was equal to 0.34331. The results show that the method is based on adaptive fuzzy neural, maintainability software performance evaluation is appropriate.
Бесплатно
Measuring the Performance of e-Primary School Systems in Bangladesh
Статья научная
The government of Bangladesh (GoB) recently have introduced the concept of Digital Bangladesh. Education is one of the most vital sectors to make the digital nation. For that reason, the GoB has started to convert primary to e-primary schools. The main objective of this study is to investigate the current Information and Communication Technology (ICT) implementation status in e-primary schools by the GoB. The study is quantitative in nature. The study also develops an ICT implementation status model from the e-primary school systems in Bangladesh. This model has identified the ICT equipment, analyzed the ICT support & equipment, given weighted to each factor and investigated the current state of ICT implementation of e-primary schools in Bangladesh. The study has taken 800 sample schools from 8 divisions to investigate the current ICT implementation status. The study suggested that before implementing the ICT they will make sure all the infrastructure of ICT is available in those primary schools.
Бесплатно
