| Paper ID |
IJIFR/V13/E9/028
|
| Author |
Peddapalli Manikanta, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
S. Manjunath Reddy, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
|
| Paper Title |
LEXIGUARD AI: A MULTILINGUAL LEGAL INTELLIGENCE AND RISK ANALYTICS ENGINE FOR AUTOMATED CONTRACT ANALYSIS
|
| Subject Category |
Computer Engineering
|
| Abstract |
The escalating complexity of legal contracts has created a structural barrier for individuals and organizations that lack access to professional legal expertise. Standard legal documents—employment agreements, non-disclosure agreements, service contracts, and commercial contracts—frequently contain high-risk clauses on liability, indemnity, and intellectual property whose implications are opaque to non-specialists. This paper presents LexiGuard AI, a modular, multilingual legal intelligence and risk analytics engine that automates contract analysis through a seven-layer processing architecture. The system accepts legal documents in PDF, DOCX, and TXT formats and processes them through sequential layers of text extraction, preprocessing, sentence segmentation, keyword-based clause classification, weighted risk scoring, SHAP-inspired word-level explainability, and interactive output presentation. Ten legally significant clause categories are supported—Liability, Indemnity, Termination, Payment, Confidentiality, Arbitration, Governing Law, Force Majeure, Intellectual Property, and Warranty—each assigned risk weights reflecting empirical legal significance. The weighted risk score (0–100) maps to four interpretive categories: Low, Medium, High, and Critical. A SHAP-inspired explainability module highlights legally sensitive terms (indemnify, not liable, waive rights, hold harmless) with importance scores that enable non-expert users to understand the reasoning behind risk assessments. Extractive summarization reduces document reading burden while preserving original legal wording. Multilingual language detection supports analysis of contracts in regional and international languages. A conversational AI assistant enables interactive clause-level querying. The system is deployed as a full-stack web application using FastAPI and Uvicorn. Experimental evaluation across four document types demonstrates reliable clause detection, interpretable risk scoring, and sub-second analysis latency for standard contracts, establishing LexiGuard AI as an accessible, explainable, and scalable foundation for democratizing legal intelligence.
|
| Keyword |
Legal document analysis; NLP; contract risk assessment; SHAP explainability; clause classification; multilingual legal AI; weighted risk scoring; FastAPI; extractive summarization
|
|
| Paper ID |
IJIFR/V13/E9/027
|
| Author |
Manchala Sarayu, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
B. Shireesha, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
|
| Paper Title |
AIRGUARDPLUS AI: AN INTELLIGENT AIR POLLUTION FORECASTING SYSTEM USING ENSEMBLE MACHINE LEARNING
|
| Subject Category |
Computer Engineering
|
| Abstract |
Air pollution has emerged as a defining environmental and public health challenge of the twenty-first century. This paper presents AirGuardPlus AI, an intelligent air quality prediction system that employs an ensemble of three independently trained machine learning models to forecast ambient Air Quality Index (AQI) values from real-time atmospheric sensor measurements. The system is built on the UCI Air Quality dataset comprising 9,471 hourly measurements recorded at a road-level monitoring station in an Italian city between March 2004 and February 2005. The ensemble architecture integrates three complementary approaches: a Random Forest regressor leveraging bagged decision trees for robust non-linear feature interaction modeling, a Gradient Boosting regressor that iteratively corrects prediction residuals through sequential ensemble construction, and a Long Short-Term Memory (LSTM) neural network that processes thirteen sensor inputs as a sequential signal to capture temporal dependencies. Predictions are produced by simple averaging across all three model outputs, reducing variance relative to any individual model. The complete system pipeline encompasses European-format data parsing, -200 sensor error code imputation using bidirectional filling, domain-informed feature engineering, 80/20 stratified data splitting, model serialization via joblib and HDF5, and a Streamlit-based web interface for real-time ensemble AQI prediction. Experimental evaluation demonstrates that the ensemble achieves an R² score of 0.932 and a Mean Absolute Error of 0.76 µg/m³, outperforming all individual component models. The system provides a practical, open-source, and locally deployable demonstration of ensemble machine learning for proactive air quality management.
|
| Keyword |
Ensemble Machine Learning; Air Quality Index; Random Forest; Gradient Boosting; LSTM; UCI Air Quality Dataset; Streamlit; Environmental Forecasting
|
|
| Paper ID |
IJIFR/V13/E9/026
|
| Author |
Dr. Meemu Dev, Assistant Professor, Department of Education, Maulana Azad National Urdu University, Hyderabad
Dr. Arun Kumar, Assistant Professor, Department of education,Indira Gandhi National Tribal University (IGNTU)
|
| Paper Title |
Impact of Mother Tongue on Academic Achievement of Tribal Students
|
| Subject Category |
Education
|
| Abstract |
Language plays a vital role in the educational development of children. For tribal students, the mother tongue serves as the primary medium through which they understand their environment, culture, and early learning experiences. However, in many Indian schools, especially in tribal regions, the medium of instruction differs from students’ native languages, creating barriers to comprehension and academic success. The present paper examines the impact of mother tongue on the academic achievement of tribal students with special reference to Indian tribal communities. The study highlights how language differences influence classroom participation, understanding of concepts, communication, confidence, and learning outcomes.
|
| Keyword |
Mother Tongue, Tribal Students, Academic Achievement, Multilingual Education, Tribal Education, Language and Learning
|
|
| Paper ID |
IJIFR/V13/E9/025
|
| Author |
Rajesh.k, VELS INSTITUTE OF SCIENCE , TECHNOLOGY& ADVANCED STUDIES (VISTAS), Chennai
Dr. R. Jeyanthi, VELS INSTITUTE OF SCIENCE ,TECHNOLOGY & ADVANCED STUDIES ( VISTAS)
|
| Paper Title |
Moral Reasoning,Ethical Awareness and Professional Self-Efficacy Among Teachers: A Systematic Review of Research from 2010 to 2026
|
| Subject Category |
EDUCATION
|
| Abstract |
Moral Reasoning, Ethical Awareness & Self-Efficacy in Teachers
Page 1 of 17
Moral Reasoning, Ethical Awareness, and Professional
Self-Efficacy Among Teachers: A Systematic Review of
Research from 2010 to 2026
ABSTRACT
This systematic review examines the relationships among moral reasoning, ethical awareness, and
professional self-efficacy in teachers, synthesising empirical and conceptual literature published between
2010 and 2026. Drawing on 52 peer-reviewed studies identified through ERIC, Scopus, PsycINFO, and
JSTOR databases using PRISMA guidelines, the review maps the theoretical foundations, measurement
approaches, development pathways, and educational outcomes associated with each construct. Findings
reveal that professional ethics and values are consistent positive predictors of teacher self-efficacy, with
particularly strong effects observed in moral and character education (ß = .54), bullying intervention, and
inclusive teaching domains. Large effect sizes (r = .61) link professional ethics to general self-efficacy.
Reflective professional learning communities, explicit ethics curricula, and structured moral deliberation
are identified as effective development mechanisms. An integrated triadic conceptual framework is
proposed, organising five constructs across three feedback loops and multiple ecological levels. Six
priority research gaps are identified, including the need for cross-cultural validation, integrated
measurement tools, and intervention trial designs. The review concludes that the moral and efficacious
dimensions of teacher professionalism are deeply and reciprocally intertwined, with significant
implications for teacher education policy and practic
|
| Keyword |
moral reasoning; ethical awareness; teacher self -efficacy; professional ethics; moral agency character education; systematic review; PRISMA
|
|
| Paper ID |
IJIFR/V13/E9/024
|
| Author |
Dr. B.PRABAKARAN, ASSOCIATE PROFESSOR OF EDUCATION, GOVERNMENT COLLEGE OF EDUCATION, PUDUKKOTTAI-622001
N. DEEPAK, M.Ed STUDENT, GOVERNMENT COLLEGE OF EDUCATION, PUDUKKOTTAI-622001
|
| Paper Title |
A STUDY ON IDENTIFYING THE REASONS FOR LACK OF INTEREST IN MATHEMATICS AMONG GOVERNMENT HIGHER SECONDARY SCHOOLS AT PUDUKKOTTAI DISTRICT
|
| Subject Category |
EDUCATION
|
| Abstract |
This paper aims to identify the various possible reasons for lack of interest in Mathematics. According to NPE (1986), “Mathematics should be visualized as the vehicle to train a child to think, reason, analyse and articulate logically. Apart from being a specific subject, it should be treated as a concomitant to any subject involving analysis and meaning”. Knowledge given in the classroom is divorced from practical life. Hence Mathematics taught in classroom is very mechanical. The subject loses its taught in an abstract, dry and uninteresting manner. Objective of the study is list out the reasons for lack of interest in Mathematics among Government higher secondary schools at Pudukkottai district. The sample of the present study consists of 300 higher secondary school students. Among them 148 are boys and 152 are girls. The survey method is followed to collect the data. The data was quantified and analysed in the form of chi-square test and percentage. Findings of the study are (i) Mathematics teacher teach the mathematical concepts fast (ii) I get stressed when mathematical concepts are to be used outside the school (iii) I do not like to play mathematical games in online and offline (iv) I do not try myself without anyone’s help to solve the problems of Mathematics. Recommendations of the study are (i) Teaching of Mathematics should be linked with familiar observations from day to day life. (ii) The best way to learn any topic is by having a basic understanding of all the concepts. Therefore, be thorough with all the basics. (iii) Students should try to understand Mathematics which is a doing subject with understandable reading the problems and related theories. (iv) Students should try to remember some tips for memorizing the difficult formulas. (v) Teachers should understand the learning difficulties encountered by each student in studying Mathematics.(vi) Teacher should create the students’ interest on maths by using modern technological aids and teaching methods. (vii) Teachers should create the students’ interest on maths by using modern technological aids and teaching methods.
|
| Keyword |
Reasons for Lack Of Interest in Mathematics, Government Higher Secondary Schools, Pudukkottai District
|
|
| Paper ID |
IJIFR/V13/E9/023
|
| Author |
Mrs. Ashwini DN, COMMUNITY INSTITUTE OF MANAGEMENT STUDIES
Dr. Manjula PS, COMMUNITY INSTITUTE OF MANAGEMENT STUDIES
Mrs. Swati N Zavar, COMMUNITY INSTITUTE OF MANAGEMENT STUDIES
Dr. Surendra H D, COMMUNITY INSTITUTE OF MANAGEMENT STUDIES
|
| Paper Title |
Smart Child Pickup System in School Trips Using OpenCV
|
| Subject Category |
Computer Science
|
| Abstract |
OpenCV (Open Source Computer Vision Library) is an open-source computer vision and machine learning library. It allows us to process images and videos, detect objects, faces that can be used in various application. This paper focus on solution to provide child safety in school trips while picking up by parents. The paper describes how opencv library helps in identifying both parent and their child in schools during trips and reduces risks in missing of child.
|
| Keyword |
Face Detection, OpenCV, Camera, Authentication, Computer Vision
|
|
| Paper ID |
IJIFR/V13/E9/022
|
| Author |
Dr. Hamsa N, Dean, School of Humanities and Social Sciences Associate Professor, Dept of Psychology Mount Carmel College, Autonomous # 58, Palace Road Bangalore 560052
|
| Paper Title |
Big Five Personality Traits and Loneliness among Siblings of Individuals with Intellectual Disability
|
| Subject Category |
Humanities and Social Sciences
|
| Abstract |
Background
The complex role of a sibling in long-term caregiving and their experienced psychological distress is often overlooked by stakeholders, interventions, rehabilitative efforts and support systems for individuals with Intellectual Disabilities.
The study explores the Big Five Personality Traits profile and loneliness using the Revised NEO Five Factor Inventory and the UCLA Loneliness Scale, version 3, including the correlation between them. 30 adult siblings of the age group 18-51 participated in the study. The data obtained were subjected to the test of normality, descriptive statistics, t-test, and correlation analysis.
Results
High levels of Neuroticism, Low Extraversion, Average Openness to Experience, Low Agreeableness, Low Conscientiousness, and Moderately High levels of Loneliness among siblings of people with intellectual disability. Results suggest there is a significant relationship between Neuroticism, Extraversion and Loneliness individually.
Conclusions
The findings provide significant insights into the traits and mental health state, having implications for developing psychological and community support systems.
|
| Keyword |
Personality, Neuroticism, Extraversion, Loneliness, Intellectual Disability, Sibling
|
|
| Paper ID |
IJIFR/V13/E9/021
|
| Author |
Dr. KANDERI SRIDEVI, Gvt. Degree College, Puttur, Andhra pradesh
|
| Paper Title |
“Mukta-Dhara” – Tagore’s Colonial Allegory and Tyranny of Nationalism
|
| Subject Category |
RESEARCH PUBLICATION
|
| Abstract |
ABSTRACT:
Tagore’s ‘Mukta-Dhara’ is a great play not only on account of its forceful presentation of a modern problem but for its highly developed technique. Mukta-Dhara is a symbolic play. Tagore’s symbolic plays are undoubtedly dramas of ideas. These ideas have been expressed through character, action, and atmosphere. The plot is well knit and the three unities of time, place and action are maintained in the play.
The play ‘Mukta-Dhar’ tells the story of how a great engineer Bibhuti builds a dam across the waters of a mountain streams which has its origin in Uttarakut Mountains and flows down to water the plains of Shiv-tarai. The royal engineer builds a dam across the waters of a mountain spring not only to prove his own scientific inventions but to enhance the political and imperial authority of his king.
The story also tells us of how the heir to the throne of that country frustrates the scientist’s dream which is devoid of all human feeling. Apart from this outline story, Mukta-Dhara has not one but several weighty themes which are interwoven with one another. Mukta-Dhara is also a symbolic play. The drama gives suggestions about some other modern problems such as the misuse of science by man, the mad pursuit of pleasure and worship of the machine, race pride and race prejudice and the regimentation of children’s minds by a slavish system of education.
“Perhaps no other play,” says K.R. Kripalani, “of Rabindranath Tagore expresses his political convictions with such directness and force”.
The drama also gives suggestions about some other modern problems such as the misuse of science by man, the mad pursuit of pleasure and worship of the machine, race pride and race prejudice, and the regimentation of children’s minds by a slavish system of education.
Tagore had thought deeply on all the questions which are connected with man’s life and happiness and he held strong convictions about them. These convictions are forcefully expressed in the ‘Mukta-Dhara’.
|
| Keyword |
Keywords: Symbolic, imperial, Prejudice, regimentation.
|
|
| Paper ID |
IJIFR/V13/E9/020
|
| Author |
S. Chandana, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
B. Shireesha, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Andhra Pradesh, India
Dr.S.Usharani, Professor, Department of Computer Applications, Viswam Engineering College,Madanapalli,Andhra Pradesh
|
| Paper Title |
INTELLIGENT EXPENSE AND BUDGET TRACKING SYSTEM USING DJANGO AND GENERATIVE AI
|
| Subject Category |
Computer Science
|
| Abstract |
In the modern digital era, effective financial management is essential for individuals and organizations to maintain economic stability and informed decision-making. Traditional methods such as manual record keeping and spreadsheet-based tracking lack real-time analytics, automation, and intelligent advisory support. This paper presents an Intelligent Expense and Budget Tracking System, a web-based application developed using the Django framework and integrated with generative artificial intelligence for enhanced financial guidance.The system enables users to record, manage, and analyze expenses through structured data storage using SQLite and Django’s Object Relational Mapping (ORM). A dynamic dashboard provides real-time insights into spending patterns, including total expenditure and category-wise distribution visualized using Chart.js. The application supports full CRUD operations, ensuring flexibility and data accuracy.A key innovation of the system is the integration of the Google Gemini AI model, which functions as an interactive financial assistant. Users can query the system using natural language to receive personalized budgeting advice and financial recommendations. This feature transforms the system from a passive tracking tool into an intelligent decision-support platform.
|
| Keyword |
Expense Tracking; Budget Management; Django Framework; Generative AI; Financial Analytics; Chart.js
|
|
| Paper ID |
IJIFR/V13/E9/019
|
| Author |
Marella Prasanth, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
B. Shireesha, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Andhra Pradesh, India
Dr.S.Usharani, Professor, Department of Computer Applications, Viswam Engineering College,Madanapalli,Andhra Pradesh
|
| Paper Title |
A Hybrid Automated Essay Scoring Framework Integrating Random Forest Regression and BERT-Based Semantic Analysis
|
| Subject Category |
Computer Science
|
| Abstract |
EssayEval AI is a comprehensive automated essay scoring system that integrates machine learning and natural language processing to deliver immediate, detailed, and actionable feedback on student-written essays. The system combines a Random Forest Regressor trained on curated linguistic features with a BERT-based transformer model to produce holistic and component-level essay evaluations. Linguistic features including word count, unique word count, vocabulary richness ratio, and average word length capture writing quality dimensions related to lexical diversity and vocabulary sophistication. The BERT model contributes a semantic quality score derived from its pretrained language understanding. The composite score is computed as a weighted combination of these two components with weights of 0.6 and 0.4 respectively. A Streamlit-based interface presents grammar, vocabulary, and coherence sub-scores alongside tailored feedback within seconds of submission. The system is designed to democratize access to high-quality writing feedback in educational environments with limited instructor availability.
|
| Keyword |
Automated Essay Scoring, Natural Language Processing, Random Forest, BERT Transformer, Formative Feedback, Streamlit, FastAPI, Machine Learning, Educational Technology
|
|
| Paper ID |
IJIFR/V13/E9/018
|
| Author |
K. Naga Sirisha, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
B. Shireesha, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Andhra Pradesh, India
Dr.S.Usharani, Professor, Department of Computer Applications, Viswam Engineering College,Madanapalli,Andhra Pradesh
|
| Paper Title |
An Intelligent CRM Framework for Small Businesses Based on Predictive Analytics and Machine Learning
|
| Subject Category |
Computer Engineering
|
| Abstract |
In the contemporary business landscape, small enterprises face significant challenges in managing customer relationships efficiently while operating with limited resources and infrastructure. This paper presents the Intelligent CRM System for Small Businesses, a web-based application developed using the Django framework, Python programming language, and SQLite database. The system is designed to centralize customer relationship management activities that are otherwise distributed across spreadsheets, email systems, and manual records.The proposed system integrates four core modules: Customer Management, Lead Tracking, Sales Recording, and Task Management. The Customer module maintains structured customer profiles, while the Lead module tracks potential business opportunities with associated value and status. The Sales module records completed transactions, enabling revenue analysis, and the Task module supports scheduling and tracking of customer-related activities.The system leverages Django’s Model-View-Template architecture and Object Relational Mapper to ensure clean separation of concerns and efficient database operations. The built-in Django administration interface provides full CRUD functionality without the need for custom frontend development, making the system immediately usable and cost-effective for small business environments.The Intelligent CRM System offers a scalable and maintainable solution that improves data consistency, operational visibility, and decision-making capabilities. It serves as a foundational platform for digital transformation in small businesses and can be extended with additional features such as analytics dashboards, role-based access control, and communication integration.
|
| Keyword |
Customer Relationship Management; Django; Web Application; Lead Tracking; Sales Management; SQLite
|
|
| Paper ID |
IJIFR/V13/E9/017
|
| Author |
Majara Ujitha, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
B. Shireesha, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Andhra Pradesh, India
Dr.S.Usharani, Professor, Department of Computer Applications, Viswam Engineering College,Madanapalli,Andhra Pradesh
|
| Paper Title |
CHURNGUARD AI: Telecom Customer Churn Prediction System Using Machine Learning
|
| Subject Category |
Computer Science
|
| Abstract |
Customer churn represents one of the most critical challenges in the telecommunications industry, where retaining existing customers is significantly more cost-effective than acquiring new ones. This paper presents ChurnGuard AI, a comprehensive end-to-end machine learning-based system designed to predict customer churn using the IBM Telco Customer Churn dataset comprising 7,043 customer records and 21 attributes. The proposed system follows a complete data science lifecycle, including data preprocessing, exploratory data analysis, feature engineering, model development, evaluation, deployment, and monitoring strategy. Four classification algorithms—Logistic Regression, Random Forest, XGBoost, and Support Vector Machine—are implemented and evaluated using five-fold stratified cross-validation. The model selection is based on the F1-score metric to effectively handle class imbalance.The best-performing model is deployed through a Flask-based web application that enables real-time churn prediction. The system provides churn probability, risk classification (Low, Medium, High), key contributing factors, and personalized retention strategies. Additionally, a production monitoring framework is designed to detect data drift and ensure long-term model reliability. The results demonstrate that the proposed system achieves strong predictive performance with practical applicability, making it a valuable decision-support tool for telecom operators to enhance customer retention strategies.
|
| Keyword |
Customer Churn Prediction; Machine Learning; Telecom Analytics; XGBoost; Predictive Analytics
|
|
| Paper ID |
IJIFR/V13/E9/016
|
| Author |
Moolinti Harshavardhan Kumar, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
B. Shireesha, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Andhra Pradesh, India
Dr.S.Usharani, Professor, Department of Computer Applications, Viswam Engineering College,Madanapalli,Andhra Pradesh
|
| Paper Title |
AI-Powered Smart Recruitment and Resume Screening System Using TF-IDF and Cosine Similarity
|
| Subject Category |
Computer Science & Engineering
|
| Abstract |
The rapid expansion of online recruitment platforms has significantly increased the volume of job applications, making manual resume screening inefficient and prone to bias. This paper presents an AI-Powered Smart Recruitment and Resume Screening System that automates candidate evaluation using Natural Language Processing and Machine Learning technique proposed system is developed using the Django framework with Python 3.10, incorporating a structured dual-role architecture for recruiters and candidates. Resume parsing is performed using the pdfplumber library, enabling accurate extraction of textual data from PDF documents. The extracted content is processed using a Term Frequency–Inverse Document Frequency (TF-IDF) vectorization approach implemented through the scikit-learn library. Candidate-job compatibility is computed using cosine similarity, producing an objective match score that ranks applicants effectively The system provides an interactive web interface built with HTML5, CSS3, and Bootstrap 5, allowing recruiters to manage job postings and evaluate candidates efficiently, while candidates can upload resumes, explore job opportunities, and track application status. The architecture follows a modular Model-View-Template design, ensuring scalability and maintainability. Experimental evaluation demonstrates that the system significantly reduces screening time, improves ranking accuracy, and minimizes subjective bias in recruitment decisions. The proposed solution offers a cost-effective and scalable alternative to traditional Applicant Tracking Systems, making it suitable for small and medium enterprises.
|
| Keyword |
Artificial Intelligence, Resume Screening; TF-IDF, Cosine Similarity, Recruitment Automation, Natural Language Processing
|
|
| Paper ID |
IJIFR/V13/E9/015
|
| Author |
Lepakshi Lathifa Bhanu, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
B. Shireesha, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Andhra Pradesh, India
Dr.S.Usharani, Professor, Department of Computer Applications, Viswam Engineering College,Madanapalli,Andhra Pradesh
|
| Paper Title |
Smart Supply Chain Tracking System A Django-Based Full-Stack Web Application for Real-Time Shipment Visibility
|
| Subject Category |
Computer Engineering
|
| Abstract |
The Smart Supply Chain Tracking System is a full-stack web application built on the Django 6.0 framework that addresses the critical challenges of transparency, accountability, and real-time visibility in modern supply chain operations. The system provides an integrated platform through which logistics managers, warehouse personnel, and procurement teams collaboratively manage the complete lifecycle of shipments — from initial consignment registration through transit updates to final delivery confirmation — alongside comprehensive supplier relationship management. Built on Django's Model-View-Template architecture with an SQLite database backend, the system implements four primary data models (Supplier, Shipment, TrackingUpdate, and Report) that collectively support end-to-end supply chain visibility. The application delivers a professional dark-themed browser interface powered by Bootstrap 5 and Chart.js, secure authentication via Django's built-in PBKDF2-SHA256 session framework, three categories of analytical reporting (shipment status, delivery delay, and supplier performance), and a public shipment tracking interface for unauthenticated consignees. All fifteen functional test cases pass, validating the correctness of the authentication, CRUD, tracking update, reporting, and public tracking modules.
|
| Keyword |
Supply Chain Management, Django Web Framework, Shipment Tracking, Logistics Visibility, MVT Architecture, SQLite, Bootstrap 5, Chart.js, Supplier Management, Real-Time Reporting, Full-Stack Web Application
|
|
| Paper ID |
IJIFR/V13/E9/014
|
| Author |
Ms. Srushti Jain, Ashoka Business School, Nashik
Dr. Manisha Shirsath, Ashoka Business School, Nashik
|
| Paper Title |
Human Capital Investment and Organizational Performance: A Study from the Management Perspective in Medium and Large-Scale Industries in Nashik City.
|
| Subject Category |
Human Resource Management
|
| Abstract |
Businesses today are under constant pressure to stay ahead, and many are realizing that the real edge lies not in machinery or capital, but in the people they employ. Yet, in practice, decisions around training, skill-building, and workforce development are often ad hoc — driven more by urgency than strategy. This study takes a closer look at how management in medium and large-scale industries in Nashik approach these decisions, and whether the investments they make actually show up in performance. The data comes directly from managerial personnel across these industries, gathered through a structured questionnaire. Using descriptive analysis and regression, the study examines what shapes human capital investment decisions and what outcomes they produce — both for individual employees and for the organization as a whole. The findings of the study are expected to show that effective investment in human capital plays a significant role in improving productivity, efficiency, and organizational growth. Nashik's industrial base is growing, but it hasn't received much attention in management research. This study is an attempt to fill that gap with evidence that is local, grounded, and practically useful.
|
| Keyword |
Human Capital Investment, Organizational Performance, Employee Performance, Management Perception, Medium and Large-Scale Industries.
|
|
| Paper ID |
IJIFR/V13/E9/013
|
| Author |
Dr.Sajitha S, Sree Narayana College for Women Kollam, Kerala
Dr.Kavitha K S, Sree Narayana College Kollam,Kerala
|
| Paper Title |
Consumer Preference Towards Unified Payments Interface (UPI) Payment Applications: An Empirical Study
|
| Subject Category |
Commerce & Finance
|
| Abstract |
Abstract
The rapid advancement of digital technology has transformed the financial transaction landscape across the world. In India, the introduction of the Unified Payments Interface (UPI) has revolutionized digital payments by enabling instant and secure transactions through mobile devices. This study examines consumer preference towards UPI payment applications in Kollam City and analyzes the factors influencing their adoption. The research is based on both primary and secondary data collected from 60 respondents through a structured questionnaire. Percentage analysis and tabular presentation were used to interpret the collected data. The findings reveal that convenience, transaction speed, ease of use, and rewards are the primary reasons for the increasing adoption of UPI applications. However, issues such as network problems, transaction failures, and security concerns still influence consumer perception. The study concludes that UPI applications play a vital role in promoting a cashless economy, but continuous improvements in security and technical reliability are necessary to sustain consumer confidence.
|
| Keyword |
UPI, Digital Payments, Consumer Preference, Financial Technology, Cashless Economy.
|
|
| Paper ID |
IJIFR/V13/E9/012
|
| Author |
Dr. B.PRABAKARAN, GOVERNMENT COLLEGE OF EDUCATION, PUDUKKOTTAI-622001
|
| Paper Title |
ASSOCIATION BETWEEN HEALTHY EATING HABITS AND ACADEMIC ACHIEVEMENT IN MATHEMATICS AMONG NINTH STANDARD STUDENTS IN PUDUKKOTTAI DISTRICT
|
| Subject Category |
EDUCATION
|
| Abstract |
Eating practices play a vital role in promoting health, longevity, and overall well-being, as they provide the essential energy required for the functioning of all cells in the human body. The objective of this study was to examine the association between healthy eating habits and academic achievement in Mathematics. A descriptive research design was adopted to investigate this relationship. The research instrument used was the Healthy Food Habits Scale, a five-point Likert-type scale consisting of 30 items with response options: Always, Most of the Time, Sometimes, Rarely, and Never. A simple random sampling technique was used to select the sample. The sample comprised 1,500 ninth-grade students selected from 60 Government and Government-Aided high and higher secondary schools in the Pudukkottai District. Data were collected using a face-to-face survey method. The normative survey method was employed for data collection. For data analysis, the chi-square test and percentage analysis were used to interpret the results. The findings revealed a statistically significant association between healthy eating habits and academic achievement in Mathematics among ninth-grade students in the Pudukkottai District.
|
| Keyword |
Association, Healthy Eating Habits, Academic Achievement in Mathematics, Ninth Standard Students and Pudukkottai District.
|
|
| Paper ID |
IJIFR/V13/E9/011
|
| Author |
Mr. Hitesh Dialani, Ashoka Business School Nashik, 422009
Ms. Srushti Jain, Ashoka Business School Nashik, 422009
|
| Paper Title |
A study on perception of non-finance students towards investment knowledge in financial market.
|
| Subject Category |
Managemnt
|
| Abstract |
This study focuses on understanding why non-finance students tend to avoid investment-related knowledge despite having basic awareness about financial markets. The main objective is to examine how factors like financial literacy, awareness, and domain perception influence avoidance behavior, and whether this behavior differs across specializations and gender. A quantitative research approach was used, and primary data was collected through a structured questionnaire from 150 MBA students in Nashik using stratified convenience sampling. The data was analyzed using statistical tools such as multiple regression, ANOVA, mediation analysis (PROCESS Macro), and chi-square test.
The results show that behavioral factors collectively have a significant impact on avoidance behavior, with financial literacy emerging as the most influential factor. The study also finds that avoidance behavior differs across specializations, indicating that academic exposure plays an important role. Mediation analysis confirms that financial literacy acts as a key link between domain perception and avoidance, while gender was found to have no significant association. Overall, the study concludes that avoidance is mainly driven by lack of understanding rather than demographic factors, highlighting the importance of improving financial literacy among students. The study highlights that merely increasing awareness is not sufficient unless it is supported by practical understanding and confidence. It also emphasizes that educational interventions should focus on simplifying financial concepts to reduce psychological barriers.
|
| Keyword |
Keywords: Financial Literacy, Avoidance Behavior, Domain Perception, Investment Knowledge, Non-Finance Students, Behavioral Finance, Mediation Analysis, Financial Awareness
|
|
| Paper ID |
IJIFR/V13/E9/010
|
| Author |
Ms. Srushti Jain, Ashoka Business School, Nashik
Dr. Manisha Shirsath, Ashoka Business School, Nashik
|
| Paper Title |
A Pilot Study on Human Capital Investment practices in Medium and Large-Scale Enterprises in Nashik.
|
| Subject Category |
Human Resource Management
|
| Abstract |
In today’s high-pressure industrial environment, the true competitive edge lies not in machinery or capital, but in the people an organization employs. Despite this, investment in training and workforce development in medium and large-scale industries is often reactive rather than strategic. This pilot study investigates the Human Capital Investment (HCI) landscape in Nashik, an industrial hub that has received limited attention in management research. Utilizing a descriptive research design, the study integrates primary and secondary data gathered through a two-stage stratified random sampling technique, which categorizes companies by scale and further divides respondents into management and employee cadres. The research is driven by four key objectives: to identify factors influencing perceptions of HCI, to analyze its impact on individual and organizational performance, to compare the differing viewpoints of management and staff, and to suggest grounded strategies for enhancing workforce outcomes. Through the application of descriptive statistics and regression analysis, the study tests the relationship between strategic investment and operational efficiency. Preliminary findings are expected to show that effective HCI significantly boosts productivity, while the pilot phase itself serves to validate the research instrument and sampling framework. Ultimately, this study attempts to fill a critical regional gap with evidence that is local, grounded, and practically useful for Nashik’s growing industrial sector.
|
| Keyword |
Human Capital Investment, Employee and Management Perception, Employee Performance, Organizational Performance, Pilot Study, Medium and Large-Scale Enterprises
|
|
| Paper ID |
IJIFR/V13/E9/009
|
| Author |
ARYA MG, JAIN UNIVERSITY BANGALOREE
ANIRUTHH G, JAIN UNIVERSITY BANGALORE
RAHUL TONY, JAIN UNIVERSITY BANGALORE
SAI SUJAL, JAIN UNIVERSITY BANGALORE
|
| Paper Title |
Contemporary Finance: Bridging Theory and Empirical Application
|
| Subject Category |
Project Centric Learning
|
| Abstract |
This research report proposes four expert-level research tracks designed for Project-Centric
Learning (PCL) in advanced finance. The overarching goal is to bridge the gap between theoretical financial
knowledge and rigorous empirical application, a gap that has widened considerably with the rapid rise of
artificial intelligence (AI), financial technology (FinTech), and alternative digital asset classes. The modern
financial ecosystem is characterised by rapid technological disruption, posing critical challenges in
governance, valuation, and risk management that traditional academic curricula have been slow to address.
|
| Keyword |
FinTech, ESG, AI/ML, Non-Stationarity, Behavioural Finance, DeFi, DCF Valuation, Project-Centric Learning, Financial Literacy, Platform Premium
|
|
| Paper ID |
IJIFR/V13/E9/008
|
| Author |
Varapana Navya, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
M. Gowthami, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
|
| Paper Title |
A Hybrid Unsupervised–Supervised Machine Learning Framework for Automated Multi-Language Code Quality Assessment: Design and Evaluation of CodeGuardian AI
|
| Subject Category |
Computer Engineering
|
| Abstract |
Maintaining high code quality across large and distributed software repositories remains a persistent challenge in modern software engineering. Traditional approaches rely on manual code reviews, rule-based linters and static analysis tools that frequently fail to capture the holistic quality of a codebase and cannot adapt to the patterns observed in real-world code. This paper presents CodeGuardian AI, an intelligent automated code quality analyzer that evaluates Python and JavaScript source files and assigns them meaningful quality ratings by combining unsupervised clustering with supervised classification in a unified machine-learning pipeline. The system collects 4,000 real-world source files (2,000 per language) from GitHub via the Search API with MD5-based deduplication, and extracts twenty static-analysis metrics per file using the Lizard library together with custom parsing logic. These metrics span structural (lines of code, cyclomatic complexity, nesting depth), documentation (comment ratio, docstring presence), style (line-length statistics, indentation consistency) and semantic (keyword density, identifier-naming quality, control-flow counts) dimensions. Features are normalized via StandardScaler, reduced to ten principal components and clustered with K-Means (k = 3); the resulting clusters are labelled Good, Average and Bad by ranking mean cyclomatic complexity. Five supervised classifiers — Random Forest, Gradient Boosting, SVM, Logistic Regression and KNN — are trained on the labelled data and achieve near-perfect accuracy. The trained Random Forest is integrated into a Streamlit web application that accepts pasted code snippets and returns instant, color-coded feedback (green for Good, yellow for Average, red for Bad). CodeGuardian AI delivers an end-to-end, multi-language, open-source solution that bridges the gap between automated metrics and human-interpretable quality judgments, lowering the barrier to rigorous quality practices for small teams, individual developers and open-source contributors.
|
| Keyword |
Code quality assessment; static analysis; unsupervised clustering; supervised classification; Streamlit deployment
|
|
| Paper ID |
IJIFR/V13/E9/007
|
| Author |
V. Kiran Kumar Reddy, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
M. Gowthami, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
|
| Paper Title |
A Computer-Vision-Augmented Web-Based Online Examination Proctoring System with Privacy-Preserving Behavioural Monitoring: Design, Implementation and Evaluation
|
| Subject Category |
Computer Engineering
|
| Abstract |
The rapid expansion of digital education and the proliferation of internet-based learning environments have created an urgent need for secure, reliable and intelligent online examination systems. Traditional examination methods that depend on physical invigilation are no longer practical when students are geographically distributed and institutions operate in hybrid or fully remote modes. This paper presents an Online Examination Proctoring System that addresses this requirement by delivering a comprehensive full-stack web application integrating computer-vision-based behavioural monitoring with conventional examination management to ensure examination integrity, automate evaluation and produce a complete audit trail. The system is implemented using Python and the Django web framework following the Model–View–Template pattern, with HTML, CSS, Bootstrap and JavaScript for the responsive frontend, OpenCV for Haar-cascade face detection, and SQLite as the relational store. Three user categories are supported: students who log in to access scheduled examinations, administrators who create and manage question banks and examination schedules, and an automated proctoring subsystem that continuously monitors examination sessions for anomalies. Real-time webcam-based facial detection identifies face-absent and multiple-face conditions; JavaScript event listeners detect browser tab-switch, focus-loss and navigation attempts; a countdown timer enforces submission upon expiry; and multi-format objective questions are auto-graded with instant result generation. Webcam frames are processed locally without streaming to external servers, preserving student privacy. The system is validated through unit, integration and user-acceptance testing, and the results confirm that the proposed design meets its functional and performance requirements while remaining deployable on commodity hardware without proprietary software or cloud dependencies, providing an accessible, privacy-respecting alternative to expensive commercial proctoring platforms.
|
| Keyword |
Online examination; AI proctoring; computer vision; Django web framework; OpenCV face detection
|
|
| Paper ID |
IJIFR/V13/E9/006
|
| Author |
V. Himateja, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
M. Gowthami, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
|
| Paper Title |
A Random-Forest-Based Disease Risk Stratification and Preventive Analytics Pipeline for Population Health Screening: Design and Evaluation of HealthRisk AI
|
| Subject Category |
Computer Engineering
|
| Abstract |
The healthcare industry continues to face persistent challenges in identifying patients at elevated risk of developing chronic or life-threatening diseases before clinical symptoms manifest in detectable form. Traditional diagnostic workflows are predominantly reactive and depend on patients seeking medical attention only after experiencing discomfort or functional impairment, an approach that is costly and produces poorer outcomes than preventive intervention strategies that intercept disease trajectories at their earliest stages. This paper presents HealthRisk AI, an end-to-end disease risk prediction and preventive analytics system that ingests patient health data containing 10,000 synthetically generated yet medically informed records covering age, body mass index, systolic and diastolic blood pressure, cholesterol, blood glucose, physical activity, smoking status and family medical history. A probabilistic data-generation framework using sigmoid-based risk scoring ensures that the synthetic dataset reflects realistic clinical distributions and risk-factor interactions. The system implements a complete machine-learning pipeline beginning with label-encoded categorical features and StandardScaler-normalized numerical features, followed by a Random Forest classifier configured with one hundred estimators trained on a stratified 80/20 train-test split, and achieves test accuracy in the range of 64–68%, consistent with the inherent uncertainty of risk-based prediction as opposed to deterministic diagnosis. Each patient record is enriched with a categorical risk label (High or Low) and a continuous probability score on a 0–100% scale, persisted to a MySQL database via SQLAlchemy. Five analytical dashboards rendered with Matplotlib and Seaborn present risk distribution, demographic risk trends, lifestyle factor analysis, biometric correlations and population-wide probability distributions, providing an actionable foundation for preventive healthcare deployment.
|
| Keyword |
Disease risk prediction; Random Forest classifier; preventive analytics; population health screening; clinical decision support
|
|
| Paper ID |
IJIFR/V13/E9/005
|
| Author |
Shaik Salma Anjum, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
Manjunath Reddy, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
|
| Paper Title |
A Value-Aware AI-Enabled Multi-Store Price Comparison Platform for E-Commerce: Design and Implementation of PriceWise
|
| Subject Category |
Computer Engineering
|
| Abstract |
The rapid expansion of electronic commerce has created an intensely competitive online retail environment in which prices for identical products frequently differ across platforms by twenty to thirty percent. Consumers seeking the best available value must visit multiple websites, record prices, evaluate seller reputations and reconcile varying product specifications — a process that is time-consuming, error-prone and frequently abandoned before completion. This paper presents PriceWise, an AI-enabled price-comparison platform that aggregates product information from multiple Indian e-commerce sources and presents it through a single, coherent interface. The platform is implemented as a full-stack web application using Python and the Django framework, with SQLite as the data store, Bootstrap 5 for responsive frontend design, and a Python-based scraping module that simulates retrieval from five major retailers — Amazon, Flipkart, Reliance Digital, Croma and Vijay Sales. A distinguishing feature of the system is its value-aware recommendation engine, which evaluates products not on price alone but on a composite value metric defined as the ratio of customer rating to price; the listing maximising this ratio is highlighted as a Smart Choice, while the lowest-priced listing is independently flagged as the Best Price. A user-specified budget threshold further enables a filtered view that respects the consumer's financial constraints. The system delivers a unified product search interface, real-time comparison across stores, automated identification of best price and best-value options, product image previews, verified ratings, direct store links, and iterative-query navigation. PriceWise demonstrates a complete end-to-end web-application workflow — from user input through data processing, database persistence and dynamic frontend rendering — and establishes a reproducible reference implementation that can be extended to live data sources, richer recommendation algorithms and personalized user experiences.
|
| Keyword |
E-Commerce; Price Comparison; Value-Aware Recommendation; Django Web Framework; Consumer Decision Support
|
|
| Paper ID |
IJIFR/V13/E9/004
|
| Author |
Shaik Nashath Rukhaiah, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
Manjunath Reddy, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
|
| Paper Title |
A Hybrid Deep Learning and Time-Series Framework for Predictive Supply Chain Disruption Analytics: Design and Evaluation of SupplyChain Sentinel AI
|
| Subject Category |
Computer Engineering
|
| Abstract |
The global supply chain ecosystem has evolved into an extraordinarily intricate network of interdependent logistics pathways, each susceptible to cascading disruptions arising from manufacturing delays, extreme weather events, geopolitical tensions, port congestion and seasonal demand fluctuations. Despite the availability of historical shipment records, many organizations continue to manage supply chain risk reactively, responding to disruptions only after they have materialized. This paper presents SupplyChain Sentinel AI, an intelligent, data-driven disruption-prediction system that combines deep learning, classical time-series forecasting and multi-dimensional risk analytics into a unified analytical platform. The system is built on a synthetically generated yet operationally realistic dataset of 2,000 shipment records annotated with carrier identity, origin and destination cities, weather conditions, traffic density, route distance, transit time goal, estimated and actual arrival timestamps, and the resulting delay in hours. Three forecasting approaches are implemented and benchmarked: a deep-learning Long Short-Term Memory (LSTM) network that captures non-linear, multi-variable relationships between shipment attributes and delay outcomes; an ARIMA statistical model that exploits temporal autocorrelation in daily average delay; and a Facebook Prophet model that decomposes the delay series into trend, seasonality and residual components. The LSTM model achieves the lowest Mean Absolute Error of 21.81 hours, outperforming ARIMA (26.06 hours) and Prophet (26.42 hours). Four analytical visualizations — a route-level risk heatmap, a high-risk route ranking, a carrier reliability assessment and a weather-impact analysis — translate model outputs into actionable business intelligence, while a MySQL backend (InspireDB) provides persistent storage and query support. The complete system is orchestrated through a modular five-stage pipeline executable as a single command, establishing a reproducible reference implementation for predictive supply chain risk management.
|
| Keyword |
Supply chain disruption; LSTM; ARIMA; Prophet; predictive analytics; risk visualization
|
|
| Paper ID |
IJIFR/V13/E9/003
|
| Author |
Chittiki Gangadhara, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
Dr. S. Usharani, Professor, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
|
| Paper Title |
Personal Finance and Investment Portfolio Tracker: A Locally Hosted Web-Based Investment Management System
|
| Subject Category |
Computer Engineering
|
| Abstract |
This paper presents a Personal Finance and Investment Portfolio Tracker, a locally hosted web application that simplifies and enhances the management of personal investments distributed across multiple accounts and brokerage platforms. Implemented in Python using the Flask framework, the system provides a centralized dashboard through which users can monitor stocks and other financial assets in a unified, real-time view. The application integrates the yfinance library to retrieve live market data — including current prices, previous closing values, 52-week highs and lows, and currency exchange rates — and stores all user, platform, account and asset records securely in a structured SQLite database following a four-level relational hierarchy. Interactive visualizations rendered with Google Charts present portfolio allocation, historical performance trends and asset-level comparisons, while a goal-tracking module allows users to set financial targets and visualise progress through dynamic indicators. Two financial simulation tools — a buy calculator and a sell calculator — let users evaluate hypothetical investment decisions before executing them by estimating profit or loss and changes in cost basis. Unlike cloud-based alternatives that depend on third-party data sharing and recurring subscriptions, the proposed system operates entirely on local hardware, ensuring data privacy and eliminating ongoing licensing cost. The application supports multiple users within a single instance, making it suitable for individual and household-level financial management. Overall, the project demonstrates the integration of modern web development, relational database management and financial analytics into an efficient, privacy-focused portfolio-tracking platform that supports real-time investment monitoring and informed decision making.
|
| Keyword |
Personal finance; portfolio tracking; Flask; SQLite; real-time market data; financial visualization
|
|
| Paper ID |
IJIFR/V13/E9/002
|
| Author |
Byrisetty Manoj Kumar, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
Dr. S. Usharani, Professor, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
|
| Paper Title |
AdOptima RL: A Deep Reinforcement Learning Framework for Real-Time Advertisement Bid Optimization
|
| Subject Category |
Computer Engineering
|
| Abstract |
The rapid evolution of programmatic advertising has introduced complex challenges in real-time bidding (RTB), where advertisers must make instantaneous and near-optimal bidding decisions in dynamic, partially observable environments. Traditional rule-based and statistical approaches struggle to adapt to fluctuations in user behaviour, market competition and platform dynamics, which frequently results in inefficient budget utilization and suboptimal campaign performance. This paper presents AdOptima RL, an intelligent advertisement optimization framework that leverages deep reinforcement learning (RL) to dynamically learn bidding strategies. The proposed system models the RTB problem as a Markov Decision Process (MDP) and integrates three RL algorithms — Deep Q-Network (DQN), Proximal Policy Optimization (PPO) and an Actor-Critic method — that collectively cover both discrete and continuous bidding action spaces. A correlation-based feature-weighting mechanism enriches the state representation by emphasising attributes that are statistically predictive of bidding outcomes, accelerating learning and improving decision quality. A per-website agent specialization strategy further allows the framework to capture domain-specific dynamics across different advertising platforms, avoiding the generalization limits of a single monolithic model. In addition to bid optimization, the framework includes a budget allocation module that simulates campaign performance using trained agents and produces data-driven recommendations for distributing advertising spend across platforms. Experimental evaluation on a real-world-inspired auction dataset demonstrates that AdOptima RL improves click-through rate and budget utilization efficiency relative to traditional methods, with PPO performing best in continuous action spaces, DQN delivering low-latency decisions in discrete spaces, and Actor-Critic offering stable convergence on long horizons. The findings highlight the potential of reinforcement learning to transform digital advertising into an adaptive, intelligent and performance-driven decision process.
|
| Keyword |
Reinforcement learning; real-time bidding; deep Q-network; proximal policy optimization; advertisement optimization
|
|
| Paper ID |
IJIFR/V13/E9/001
|
| Author |
Shaik Muqthiyar, Student, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
S. Manjunath Reddy, Assistant Professor, Department of Computer Applications, Viswam Engineering College, Madanapalle, Andhra Pradesh, India
|
| Paper Title |
A Lightweight Transformer Framework for English-to-Hindi Neural Machine Translation: Design, Training and Web-Based Deployment of IndicTrans AI
|
| Subject Category |
Computer Engineering
|
| Abstract |
IndicTrans AI is a Transformer-based neural machine translation system engineered to bridge the communication gap between English and Hindi, one of the most widely spoken languages of the Indian subcontinent. With more than half a billion Hindi speakers across India and the global diaspora, the demand for accurate, efficient and computationally feasible automated translation systems remains substantial. The proposed system implements a lightweight yet structurally complete encoder-decoder Transformer architecture, trained on curated English-to-Hindi sentence pairs and deployed through a Streamlit web interface that makes neural translation accessible to non-technical users. The system is implemented entirely in Python using PyTorch as the deep-learning framework. A SimpleTransformer model class encapsulates two embedding layers, an nn.Transformer encoder-decoder block configured with a model dimension of sixty-four, two attention heads and single encoder and decoder layers, and a linear output projection. The model is trained using the Adam optimizer with cross-entropy loss over fifty epochs, while source and target vocabularies are built from the corpus through word-level tokenization with reserved indices for padding and unknown tokens. Trained model weights and vocabulary mappings are persisted using PyTorch state-dictionary serialization and Python's pickle module, enabling efficient reuse across inference sessions. The Streamlit-based web application provides an intuitive two-column interface in which users enter English text and receive the Hindi translation in real time; output words are reconstructed from predicted token indices via a reverse Hindi vocabulary mapping. IndicTrans AI demonstrates the feasibility of Transformer-based machine translation at a compact scale, providing a foundation for extending coverage to additional Indian regional languages and incorporating more sophisticated training regimes and larger multilingual corpora.
|
| Keyword |
Neural machine translation; Transformer; English to Hindi; PyTorch; Streamlit
|
|
|
|