Journal of Science Engineering Technology and Management Science

The study suggests an automated way of preventing bogus job postings online that uses categorisation techniques based on machine learning. To determine the most effective model for identifying job scams, the output of multiple classifiers was compared. In order to verify false internet postings, these classifiers are used. In the midst of several other ads, it aids in identifying fraudulent job listings. The two fundamental categories of classifiers considered for the purpose

Keywords :

Author : Shreeja.V, Ashitha.P,R.Raj Kumar

Title : A COMPARATIVE STUDY OF FAKE JOB POST PREDICTION USING DIFFERENT DATA MINING TECHNIQUES

Volume/Issue : 2025;02(10)

Page No : 1-3

The rapid advancements in natural language generation have introduced powerful tools capable of shaping public opinion on social media platforms. Enhanced language modeling techniques have significantly improved the generative capabilities of deep neural networks, enabling them to produce highly realistic and contextually accurate text. This progress has led to the emergence of sophisticated text-generative models that can be exploited by adversaries to power social bots, creating convincing deepfake posts that influence public discourse. To combat this growing threat, the development of robust and accurate detection methods for identifying machine-generated content is essential. In response, this study focuses on the detection of deepfake tweets on platforms such as Twitter. A deep learning-based approach is proposed, employing a Convolutional Neural Network (CNN) architecture in combination with FastText word embeddings to classify tweets as either human-generated or botgenerated. The model is trained and evaluated using the publicly available Tweepfake dataset. To validate the effectiveness of the proposed method, it is benchmarked against several traditional machine learning models utilizing features like Term Frequency (TF), Term Frequency– Inverse Document Frequency (TF-IDF), FastText, and FastText subword embeddings. Additionally, comparisons are made with other deep learning architectures, including Long Short-Term Memory (LSTM) and hybrid CNN-LSTM models.

Keywords : Deepfake detection, social bots, text generation, CNN, FastText, Tweepfake dataset, machine-generated text, Twitter, natural language generation, fake tweets, TF-IDF, LSTM, CNNLSTM, social media manipulation, bot detection

Author : Prathima Patnaik,M.koushik Reddy, K.Bala Krishna

Title : DEEPFAKE DETECTION ON SOCIAL MEDIA LEVERAGING DEEP LEARNING AND FASTTEXT EMBEDDINGS FOR IDENTIFYING MACHINE-GENERATED TWEETS

Volume/Issue : 2025;02(10)

Page No : 4-8

In-personbankingisstill an important part of financial services around the world. Hybrid bank branches with service robots can improve efficiency and reduce operating costs. An efficient autonomous Know-Your-Customer (KYC) is required for hybrid banking. In this paper, an automated deep learning based framework for interbank KYC in robot-based cyber-physical banking is proposed. A deep biometric architecture was used to model the customer’s KYC and anonymize the collected visual data to ensure the customer’s privacy. The symmetric-asymmetric encryptiondecryption module in addition to the blockchain network was used for secure and decentralized transmission and validation of the biometric information. Ahigh-capacity fragile watermarking algorithm based on the integer-to-integer discrete wavelet transform in combination with the Z6 and A6 lattice vector quantization for the secure transmission and storage of in-person banking documents is also proposed. The proposed framework was simulated and validated using a Pepper humanoid robot for the automated biometric-based collection of handwritten bank checks from customers adhering to COVID-19 pandemic safety guidelines. The biometric information of bank customers such as fingerprint and name is embedded as a watermark in the related bank documents using the proposed framework. The results show that the proposed security protection framework can embed more biometric data in bank documents in comparison with similar algorithms. Furthermore, the quality of the secured bank documents is 20% higher in comparison with other proposed algorithms.

Keywords : In-person banking, Hybrid bank branches, Service robots, Efficiency, Operating costs, Autonomous KYC, Deep learning framework, Biometric architecture, Customer privacy, Encryption, Blockchain network, Secure transmission

Author : M.Pragnika, B.Akshara, P.Shailaja

Title : IORT-Enabled Banking: A Cyber-Physical Approach to Customer Management

Volume/Issue : 2025;02(10)

Page No : 9-12

For driver state classification, the suggested system employed the k-NN approach. It has not before been explored in the context of a camera-based driver sleepiness detection employing blink features, to the best of our knowledge. Steering behavior, EEG measurements, and facial traits are examples of existing kNN-based techniques. The research looks into the viability of a drowsiness classification system based on blink features collected using an EOG. The author attained a promising classification accuracy, demonstrating the utility of a k-NN classifier combined with blink-based features. When a highdimensional feature space is available, the k-NN model requires a set of acceptable features as a basis for classification. The accessible data becomes scarce as the number of alternative configurations increases, according to the "curse of dimensionality" phenomenon. grows. As a result, one goal of this research is to discover an appropriate set of significant traits. Wrapper approaches are the most commonly utilized feature selection strategies in this work. Wrapper approaches choose feature subsets based on their predictive value during the classification phase. As a result, because it directly evaluates classification performance, this method can take into account dependencies between the feature subset and the classifier.

Keywords :

Author : P.Usha Sree, M.Vasuki Deepika, J.Srividya

Title : PREDICTING DRIVER DROWSINESS USING KNN ALGORITHM

Volume/Issue : 2025;02(10)

Page No : 13-17

With technological advancements, data and its analysis have evolved beyond simple values and attributes scattered across spreadsheets, becoming a powerful catalyst for transformation across numerous fields. However, data corruption, often stemming from unethical or illegal activities, has emerged as a serious challenge, highlighting the urgent need for effective methods to detect and clearly identify corrupted data within datasets. Identifying and recovering corrupted data is a complex task that demands significant attention, as overlooking it during early stages can lead to major complications in subsequent machine learning or deep learning processes. In this work, we introduce PAACDA (Proximity-based Adamic Adar Corruption Detection Algorithm) and present consolidated results with a specific emphasis on detecting corrupted data rather than merely identifying outliers. Existing state-of-the-art models like Isolation Forest and DBSCAN (Density-Based Spatial Clustering of Applications with Noise) depend heavily on meticulous parameter tuning to achieve high accuracy and recall; nevertheless, they still exhibit considerable uncertainty when it comes to handling corrupted data. The present study focuses on addressing niche performance limitations of various unsupervised learning algorithms when applied to both linear and clustered corrupted datasets.

Keywords : exhibit, meticulous, considerable, learning

Author : M.Shrijani, BP Gitika Krishna, Dr.S.Jayanth

Title : PAACDA: COMPREHENSIVE DATA CORRUPTION DETECTION ALGORITHM

Volume/Issue : 2025;02(10)

Page No : 18-22

This project shows a network of wearable health sensors that can connect to the Internet of Things (IoT) for safety and health care purposes. The system uses wearable sensors to keep an eye on things like temperature, vibration, and heart rate in the body and the environment. An Arduino module processes the data from the sensors, an LCD shows it, and an intelligent IoT gateway sends it to a server. The gateway connects to both a web server on the local network and the cloud for processing, storing, and displaying data. The system lets you watch patients in real time, finds unusual health problems, and sends out alerts quickly to make patient care and hospital staff safety better.

Keywords : IoT, health, SPO2, heart rate, LCD, DHT11, safety, and protection.

Author : Raniya Zainab,V. Mahender

Title : SMART HEALTH GUARDIAN: IOT-ENABLED REAL-TIME HEALTH AND SAFETY MONITORING

Volume/Issue : 2025;02(10)

Page No : 23-31

By utilizing IoT and advanced wireless communication technologies, the "LoRa-Based Soldier Security System Integrated with IoT" seeks to improve troops' safety and real-time surveillance while they are in the field. The two main components of the system are the transmitter (Tx) and the receiver (Rx). An LCD display, a buzzer, a LoRa transmitter module, a keypad emergency button, and essential sensors like SpO2 and cardiac sensors are all integrated into the Tx unit. The Rx unit has an LCD display, a buzzer, a LoRa receiver module, and Internet of Things integration for remote monitoring. On the Tx side, data from the sensors is continuously tracked, including vital health indicators. Both the Tx and Rx devices send out a warning through buzzers and LCD screens if any of the parameters exceed predetermined threshold levels or if the emergency button is touched. The alarm is sent to an Internet of Things server at the same time so that higher authorities are notified in real time. An ESP32 is used on the Rx side of the system for Internet of Things connectivity, while an ESP32 is used on the Tx side for sensor data collection. This project provides a long-range, dependable, low-power communication solution that will improve overall military safety and operational efficiency while guaranteeing prompt replies in emergency scenarios.

Keywords : Lo-Ra, IoT, Soldier, Security, Spo2, ESP32 Controller.

Author : Bangari Nikhilvaas,V. Mahender

Title : WIRELESS LORA COMMUNICATION BASED SOLDIER SECURITY ALERT SYSTEM

Volume/Issue : 2025;02(10)

Page No : 32-41

The rapid evolution of electric vehicles (EVs) has necessitated advanced power conversion and battery management systems to enhance efficiency, reliability, and compactness. This project presents an integrated charging system for power and auxiliary batteries in EVs, employing a Dual Active Bridge (DAB) converter topology to achieve high-efficiency bidirectional power transfer between high-voltage (HV) and low-voltage (LV) batteries. The proposed system incorporates a shared magnetic coupler and compensation network for the wireless charging system (WCS) and the auxiliary power module (APM), minimizing component count and overall system weight. The design facilitates seamless operation in multiple modes, enabling both grid-to-vehicle and vehicle-toauxiliary battery charging. Simulation and analysis are conducted using MATLAB/Simulink to evaluate steady-state and dynamic performance parameters, such as voltage regulation, current response, and Zero Voltage Switching (ZVS) efficiency. The integrated system demonstrates superior performance with reduced cost, improved power density, and enhanced energy management, thereby representing a viable solution for next-generation EV charging infrastructure.

Keywords : Electric Vehicle (EV), Dual Active Bridge (DAB), Wireless Charging System (WCS), Auxiliary Power Module (APM), Bidirectional Power Flow, Zero Voltage Switching (ZVS), MATLAB/Simulink

Author : J. Jaya Krishna, M. Sridhar, K. Dasu, A. Thirupathireddy, V. Prasanna Rakesh,Dr. Rushi Santhosh Singh Thakur

Title : INTEGRATED CHARGING SYSTEM FOR POWER AND AUXILIARY BATTERIES IN ELECTRIC VEHICLES USING DUAL ACTIVE BRIDGE CONVERTER

Volume/Issue : 2025;02(10)

Page No : 42-53

Failure of facial recognition and authentication system may lead to several unlawful activities. The current facial recognition systems are vulnerable to different biometric attacks. This research focuses on morphing attack detection. This research proposes a robust detection mechanism that can deal with variation in age, illumination, eye and head gears. A deep learning-based feature extractor along with a classifier is adopted. Additionally, image enhancement and feature combination are proposed to augment the detection results. A versatile dataset is also developed that contains Morph-2 and Morph-3 images, created by sophisticated tools with manual intervention. Morph-3 images can give more realistic appearance and hence difficult to detect. Moreover, Morph-3 images are not considered in the literature before. Professional morphing software depicts more realistic morph attack scenario as compared to the morphs generated in the previous work from free programs and code scripts. Eight face databases are used for creation of morphs to encompass the variation. These databases are Celebrity2000, Extended Yale, FEI, FGNET, GT-DB, MULTIPIE, FERET and FRLL. Results are investigated using multiple experimental setups and it is concluded that the proposed methodology gives promising results.

Keywords : Morph-2 and Morph-3 images, Celebrity2000, Extended Yale, FEI, FGNET, GT-DB, MULTIPIE, FERET and FRLL

Author : Mr.S.Vamshi Krushna, Voluru Saranya, Kotha Sneha, N Ajay Kumar Reddy, Bukya Rajkumar

Title : GENERATION AND DETECTION OF FACE MORPHING ATTACKS

Volume/Issue : 2025;02(10)

Page No : 54-68

The project "Big Data Job Analysis" delves into the dynamics of the contemporary job market, specifically focusing on positions related to big data. Through a comprehensive analysis of job postings, skill requirements, and industry trends, this study aims to provide insights into the evolving landscape of big data jobs. The findings contribute valuable information for job seekers, employers, and educators, shedding light on the skills and qualifications in demand for successful

Keywords :

Author : Mr.S.Vamshi Krushna, Vanam Niharika, Thokati Udaykiran, Kayithi Raj Kumar, Marella Naga Sai

Title : BIGDATA JOB ANALYSIS

Volume/Issue : 2025;02(10)

Page No : 69-78

The project on "Deep Learning Approach for Multimodal Biometric Recognition System Based on Fusion of Iris, Face, and Finger Vein Traits" presents an advanced biometric authentication system that integrates multiple modalities. Leveraging deep learning techniques, the system combines iris, face, and finger vein traits to enhance the accuracy and security of biometric recognition. Biometric recognition systems have become indispensable in ensuring security and authentication across various domains. However, the reliability and accuracy of these systems are often challenged by environmental factors, intra-class variations, and attempts at spoofing. To address these challenges, this project proposes a novel deep learning-based approach for multimodal biometric recognition, integrating iris, face, and finger vein traits. The proposed system leverages the complementary nature of multiple biometric modalities to enhance recognition accuracy and robustness. Deep learning techniques, particularly convolutional neural networks (CNNs), are employed for feature extraction and fusion from each modality. The fusion process aims to combine the distinctive characteristics of iris, face, and finger vein traits to create a comprehensive and discriminative biometric template.

Keywords : Biometric recognition systems, convolutional neural networks (CNNs), Finger Vein Traits, Deep learning.

Author : Mrs.K.Keerthi Reddy, Mittakola Gnaneshwar, Boddu Amulya, K Sukrutha, Bitla Vijay

Title : DEEPLEARNING APPROACH FOR MULTIMODAL BIOMETRIC RECOGNITION SYSTEM BASED ON FUSION OF IRIS, FACE, AND FINGER VEINTRAITS

Volume/Issue : 2025;02(10)

Page No : 79-89

This study presents the application of a Multi-Objective Goal Programming (MOGP) model with timeline analysis to optimize multiple, conflicting objectives over a set time period. The model focuses on minimizing costs, optimizing resource utilization, achieving high service levels, and reducing environmental impact across four time periods (Q1 to Q4). The results indicate that while some objectives, such as cost minimization and service level achievement, faced minor deviations from the target goals, overall performance remained within an acceptable range. Specifically, cost minimization showed a deviation of -$20,000 in Q1, but the goal was overachieved by $10,000 in Q3. Resource utilization efficiency consistently met or exceeded targets, while environmental impact targets experienced slight deviations, peaking at 2 tons in Q1 but stabilizing in Q4. The sensitivity analysis of goal weights demonstrated the model's robustness, with changes in goal weights leading to variations in resource allocation and prioritization, allowing for flexible decision-making in response to evolving priorities. The results highlight the effectiveness of the MOGP model with timeline analysis in real-world decision-making problems, balancing multiple objectives while adapting to changing goals over time

Keywords : Multi-Objective Goal Programming, Timeline Analysis, Cost Minimization, Resource Utilization, Service Level Achievement.

Author : Nagireddy Rajender Reddy,T.Yugandhar

Title : DEVELOPMENTS IN MULTI-OBJECTIVE MODEL WITH A TIMELINE ANALYSIS USING GOAL PROGRAMMING

Volume/Issue : 2025;02(10)

Page No : 90-95