Articles

Access the latest knowledge in applied science, electrical engineering, computer science and information technology, education, and health.

Advanced Search
Filter Icon

Filters article

Years

FAQ Arrow
Single Year
Range Year
0
0

Source Title

FAQ Arrow

Authors

FAQ Arrow

30,411 Article Results

Large language models and retrieval-augmented generation-based chatbot for adolescent mental health

10.11591/ijaas.v14.i3.pp849-858
Andi Riansyah , Imam Much Ibnu Subroto , Intan Nur'aini , Ratna Supradewi , Suyanto Suyanto
Access to fast and efficient information is crucial in today's digital era, especially for teenagers in obtaining mental health services. The manual method used by Youth Information and Counselling Centre (PIK R) to provide mental health information requires significant time and effort. This research presents an AI-based solution by developing a chatbot system using retrieval-augmented generation (RAG) and large language models (LLM). This chatbot is designed to provide accurate and effective mental health information for teenagers throughout the day. An analysis of a dataset consisting of articles on teenage mental health and data from the Alodokter website was used as the basis for the development of this chatbot. The research results show that the chatbot is capable of providing relevant and accurate information, with evaluations using the recall-oriented understudy for gisting evaluation (ROUGE) score method yielding an average of ROUGE-1 with a precision of 87.8%, recall of 83.0%, and F1-measure of 84.0%; ROUGE-2 with a precision of 82.8%, recall of 76.8%, and F1-measure of 78.2%; and ROUGE-L with a precision of 88.0%, recall of 82.6%, and F1-measure of 83.4%. These findings indicate the potential use of chatbots as an effective tool to support the mental health of adolescents.
Volume: 14
Issue: 3
Page: 849-858
Publish at: 2025-09-01

Optimizing retail systems: using big data and power business intelligence for performance insights

10.11591/ijaas.v14.i3.pp945-954
Huu Dang Quoc , Ha Le Viet
In the rapid development of information technology, using enterprise data to support timely management decisions is crucial in helping businesses operate effectively and improve competitiveness. This study uses Microsoft power business intelligence (MPBI) to analyze data in retail systems, allowing managers to grasp the business situation in real time, track advanced sales, optimize inventory control, and analyze customer behavior and supply chain visibility. From the data generated by the business, the study uses the streaming extract transform load (ETL) model to support real-time data aggregation, then converts to the MPBI data visualization system to convert data into visual charts, helping businesses easily monitor, track, analyze, and make decisions to promote business activities. The study proposes a data structure to organize retail information storage. It proposes a system of calculation formulas and data synthesis, making integrate and convert tabular data into visual charts. Through analysis of real data from the LH83 retail system, the study shows the feasibility of implementing a data visualization system and the difficulties encountered when businesses want to deploy this model.
Volume: 14
Issue: 3
Page: 945-954
Publish at: 2025-09-01

Haystack-based Facebook’s data storage architecture: store, directory, and cache

10.11591/ijaas.v14.i3.pp671-681
Tole Sutikno , Ahmad Heryanto , Laksana Talenta Ahmad
Haystack is Facebook's unique way of managing large amounts of user-generated content like photos. The architecture prioritizes performance, reliability, and scalability to overcome network-attached storage system bottlenecks. Haystack speeds data access and ensures data integrity during hardware failures by using physical and logical volumes. This study examines the architecture of Facebook's Haystack data storage system and its effects on scalability and efficiency in handling large photo data. According to the study, the store, directory, and cache functions work together to reduce input/output (I/O) operations and improve metadata processing, which traditional network-attached storage systems cannot do. Haystack manages massive photo data storage and retrieval, solving network-attached storage (NAS) limitations. It balances throughput and latency by minimizing disk operations and optimizing metadata processing. Each store, directory, and cache contribute to this ecosystem. The Haystack architecture reduces disk operations and metadata processing bottlenecks with distributed caching. A cache allows instant access to frequently requested images and balances read and write operations across the system. We should study advanced storage system architectures based on Facebook's Haystack architecture. This could involve investigating faster metadata processing algorithms, using artificial intelligence (AI) to improve fault detection and repair systems, and assessing the economic impact of distributed caches.
Volume: 14
Issue: 3
Page: 671-681
Publish at: 2025-09-01

Test rig development for load test of pipe saddle support

10.11591/ijaas.v14.i3.pp886-893
Muhammad Arif Rayhan , Mohd Shukri Yob , Mohd Juzaila Abd Latif , Ojo Kurdi , Fudhail Abdul Munir
Pipe saddle support is a structure commonly used to support horizontal steel pipe. It prevents direct contact between the pipe and the support. Pipe saddle support can experience displacement due to pipe movement and insufficient stress analysis. Given these concerns, conducting a load test is essential to determine the stress on pipe saddle supports. However, a universal testing machine (UTM) is not suitable for this purpose due to the size limitation. Therefore, this study proposed a test rig setup for the pipe saddle support load test. The test rig consists of a portal frame secured by an underground locking system featuring a strong floor. Additionally, an actual pipe is utilized to replicate actual loading conditions on the pipe saddle support. The applied load is measured using a load cell, with a custom-designed bracket to ensure precise load transfer. Finally, the pipe saddle support specimen is bolted to a base support to maintain stability during the load test. Stress analysis using finite element analysis (FEA) demonstrated that the test rig is suitable for conducting load tests on the specimens with a maximum force of 80 kN. FEA confirmed that the test rig operates within a safety factor of 1.3.
Volume: 14
Issue: 3
Page: 886-893
Publish at: 2025-09-01

E-commerce waste management: a systematic review

10.11591/ijaas.v14.i3.pp817-827
Mohd. Suhaimi Shamsuddin , Noor Fadhiha Mokhtar , Safiek Mokhlis , Zuha Rosufila Abu Hasan , NajdahAbd Aziz , Mohamad Nizam Yusof
This paper reviews literature on e-commerce waste management issues and challenges, focusing on potential improvements in Malaysia. It analyzes various sources, including Scopus, Web of Science (WoS), and Google Scholar (GS), using thematic and content analysis based on preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines. The review highlights the surge in packaging and electronic waste due to increased e-commerce activity. In response, Malaysia has introduced policies promoting sustainable practices, such as eco-friendly packaging, e-waste regulations, and circular economy (CE) principles. Growing consumer awareness has also driven demand for sustainable e-commerce options. However, the key challenge is to reduce waste generation rather than just managing it. Achieving this will require significant efforts to minimize excessive manufacturing and packaging. The review aims to provide insights for stakeholders to support effective waste management and foster sustainability in the e-commerce sector.
Volume: 14
Issue: 3
Page: 817-827
Publish at: 2025-09-01

Pitch extraction using discrete cosine transform based power spectrum method in noisy speech

10.11591/ijaas.v14.i3.pp955-965
Humaira Sunzida , Nargis Parvin , Jafrin Akhter Jeba , Sulin Chi , Md. Shiplu Ali , Moinur Rahman , Md. Saifur Rahman
The pitch period is a key component of many speech analysis research projects. In real-world applications, voice data is frequently gathered in noisy surround- ings, therefore algorithms must be able to manage background noise well in order to estimate pitch accurately. Despite advancements, many state-of–the-art algorithms struggle to deliver adequate results when faced with low signal-to- noise ratios (SNRs) in processing noisy speech signals. This research proposes an effective concept specifically designed for speech processing applications, particularly in noisy conditions. To achieve this goal, we introduce a fundamen- tal frequency extraction algorithm designed to tolerate non-stationary changes in the amplitude and frequency of the input signal. In order to improve the extrac- tion accuracy, we also use a cumulative power spectrum (CPS) based on discrete cosine transform (DCT) rather than conventional power spectrum. We enhance extraction accuracy of our method by utilizing shorter sub-frames of the input signal to mitigate the noise characteristics present in speech signals. According to the experimental results, our proposed technique demonstrates superior per- formance in noisy conditions compared to other existing state-of-the-art meth- ods without utilizing any kind of post-processing techniques.
Volume: 14
Issue: 3
Page: 955-965
Publish at: 2025-09-01

Artificial neural network based sensorless position estimation and direct torque control for stepper motor

10.11591/ijaas.v14.i3.pp702-710
Nagasridhar Arise , Thiruveedula Madhu Babu , Srinidhi Gollapudi , Tarun Kumar Dommeti , Abhishek Kummari , Mahith Shambukari
This study describes and illustrates how sensorless location estimation is achieved through the application of artificial neural network (ANN) control. Control stepper motor torque directly. Using stepper motors directly leads to a lot of problems; therefore, automated control systems are now commonly preferred. Stepper motors have several drawbacks when used directly, including the potential for steps to occasionally be missing while the motors are running. When physical sensors are not available, the proposed method estimates rotor position and speed using electrical signals and ANN algorithms. Simulation and experiment results demonstrate accurate position estimation (±1.5°) and efficient torque control. The sensorless direct torque control (DTC)-ANN approach increases the performance, reliability, and cost of stepper motors in robotics, computer numerical control (CNC) machines, and 3D printing.
Volume: 14
Issue: 3
Page: 702-710
Publish at: 2025-09-01

Solar photovoltaic based cascaded multilevel inverter with 33-levels using phase opposition disposition control method

10.11591/ijaas.v14.i3.pp928-935
Chandolu Sai Deepak , Madhu Babu Thiruveedula , Bandari Rahul Teja , Supe Gowtham , Sthambhampally Vivek , Panuganti Yeshwanth Kumar
A cascaded multilevel inverter (MLI) tailored for photovoltaic (PV) networks, aiming to improve power quality and support transformer-less operation. The symmetric MLI design is selected for its effectiveness in minimizing harmonics and enhancing fault tolerance in high-power scenarios, where the use of power semiconductor converters can introduce complications. The proposed inverter configuration achieves thirty-three voltage levels, optimizing power quality while using insulated gate bipolar transistor (IGBT) semiconductor switches. The phase opposition disposition (POD) control method is applied to trigger necessary switching signals for the inverter's components. To ensure high output voltage for the MLI, a boost converter is employed, and the overall system is tested with an R load. The effectiveness of the design is validated through MATLAB/Simulink simulations, which demonstrate a notable reduction in total harmonic distortion (THD). 
Volume: 14
Issue: 3
Page: 928-935
Publish at: 2025-09-01

Determination of soil salinization by hyperspectral remote sensing in the Shirvan Plain

10.11591/ijaas.v14.i3.pp662-670
Sahib Shukurov Khudaverdi , Aygun Ismayilova Azer , Ramil Sadigov Ali , Maya Karimova Javanshir , Turkan Hasanova Allahverdi , Gunel Asgarova Farhad
The determination of soil salinization in the Shirvan Plain, considered the main agricultural zone of Azerbaijan, negatively affects the productivity of agricultural crops. Based on 10 m Sentinel-2 images on Google Earth Engine platforms and by examining SI1, green-red band normalized difference vegetation index (GRNDVI), green normalized difference vegetation index (GNDVI), normalized difference vegetation index (NDVI), and difference vegetation index of the environment (DVI), four remote sensing salinity monitoring index models, S1DI1, S1DI2, S1DI3, and S1DI4, were constructed to extract soil salinity information in the Shirvan Plain in combination with the measured electrical conductivity. The results show that the overall classification accuracy of S1DI1 (SI1-GRNDVI), S1DI2 (SI1-GNDVI), S1DI3 (SI1-NDVI), and S1DI4 (SI1-DVI) models for salinity monitoring are 82.35%, 83.10%, 81.96%, and 79.25%, respectively.
Volume: 14
Issue: 3
Page: 662-670
Publish at: 2025-09-01

Impulse buying behavior in mobile commerce: a partial least squares structural equation modeling analysis

10.11591/ijaas.v14.i3.pp761-772
Hery Hery , Winnie Veronica , Andree E. Widjaja , Calandra Alencia Haryani , Riswan E. Tarigan
Efficient online transactions now thrive through websites (e-commerce) and mobile apps (m-commerce). With the growth of m-commerce, marketers aim to boost profits by understanding impulsive buying behavior. This study investigates factors influencing online impulse buying (OIB) in m-commerce by analyzing key variables. Data were gathered via questionnaires from 449 Indonesian consumers who had made digital payments and impulsive purchases using smartphones. The framework includes sales promotion (SP), attractive advertising (EA), and mobile digital payment systems (MDPS) as situational factors; hedonic shopping motivation (HSM) as a motivational factor; and impulsiveness (I) and smartphone addiction (SA) as personal traits. Analysis used partial least squares structural equation modeling (PLS-SEM), with gender, income, and smartphone usage time as control variables. Results show that EA, MDPS, HSM, I, and SA significantly influence OIB, while SP does not. For consumer segmentation, t-distributed stochastic neighbor embedding (t-SNE) outperformed ISOmap and principal component analysis (PCA), achieving a silhouette score of 0.72. A paired t-test (p<0.01) confirmed t-SNE’s superior clustering accuracy. These findings reveal that t-SNE better captures consumer segmentation patterns, helping businesses refine marketing strategies and deepen their understanding of psychological drivers behind impulsive m-commerce purchases.
Volume: 14
Issue: 3
Page: 761-772
Publish at: 2025-09-01

ToLatin application acceptability evaluation to support Balinese script transliteration learning

10.11591/ijaas.v14.i3.pp804-816
Luh Joni Erawati Dewi , Gede Indrawan , I Made Agus Oka Gunawan , I Wayan Sutaya , Sariyasa Sariyasa
This work supported Indonesia's research focus area on information and communication technology (ICT) content improvement for information data on various forms of local wisdom. As one of the various forms of local wisdom, the Balinese script was supported by the ToLatin application that transliterates Balinese script into Latin text. It has been used to support Balinese script learning at the high school level in Buleleng Regency, Bali, Indonesia. To determine the acceptability of this application, which had not been studied before, a user acceptance evaluation was conducted using a combination of acceptance variables from the technology acceptance model (TAM) and success variables from DeLone & McLean. This study used a quantitative method with data collection through questionnaires from 385 respondents. The data analysis used the importance-performance analysis (IPA) method through suitability, gap, and quadrant analyses. The study results indicated that the acceptability of ToLatin could be more optimal. The suitability analysis revealed an average score of 87.91%, indicating the need for improvement in system quality, particularly the innovative indicator (SysQ3), based on the quadrant analysis. The gap analysis revealed an average score of -0.54 from 7 acceptance variables, indicating the need to improve system performance to meet user expectations.
Volume: 14
Issue: 3
Page: 804-816
Publish at: 2025-09-01

An innovative design and development of multilevel inverter for a wind energy conversion system

10.11591/ijaas.v14.i3.pp751-760
Rosaiah Mudigondla , Thiruveedula Madhu Babu , Supriya Dachepalli , Anudeep Panjula , Md Yousuf Ali , Bakam Anirudh
The drawbacks of fossil fuel-based energy sources, including high costs, pollution, scarcity, and environmental damage, highlight how urgent it is to switch to renewable energy sources. Multilevel inverters (MLIs) are currently required for the production of wind electricity. In this research, to get a reduced harmonic distortion, use 31-level inverter based on shifted carrier-pulse width modulation (SC-PWM) is developed for wind power generation using MATLAB/Simulink. It aids in minimizing the total harmonic distortion (THD) to 3.20, and the output voltage is enhanced by the suggested MLI. Wind energy extraction is optimized by combining with a proportional integral derivative (PID) control system. MATLAB/Simulink has been used to make sure the MLI structure and look into the implementation of wind energy conversion systems using a permanent magnet synchronous generator (PMSG). In order to show that the suggested inverter architecture improves power conversion efficiency and stability in renewable energy systems, the study also examines power efficiency, system dependability, and the viability of large-scale applications. Additionally, the study investigates grid integration, modulation strategies, and switching losses to guarantee increased sustainability, dependability, and efficiency in wind energy applications while lowering operating costs.
Volume: 14
Issue: 3
Page: 751-760
Publish at: 2025-09-01

Gated dilated causal convolution-based encoder-decoder network for IoT intrusion detection

10.11591/ijape.v14.i3.pp722-732
Aarthi Gopalakrishnan , Sharon Priya Surendran , Aisha Banu Wahab
The internet of things (IoT) is perhaps the greatest modern development, as it affects our daily lives and is rapidly expanding in its application zones. The IoT is used in everyday activities, so security is more crucial because intrusion detection will introduce and eliminate attacks. In this paper, a novel deep learning based intrusion detection technique (DEBIT) has been proposed that detects the intrusion using deep learning techniques efficiently. Initially, the data from IoT user is preprocessed and classified using the novel gated dilated casual convolution based encoder-decoder (GDCC-ED) method, which classifies the data into attack and non-attack. The proposed DEBIT framework has been assessed using a MATLAB simulator. The performance of the proposed DEBIT framework has been assessed based on specific parameters, including recall, detection rate, accuracy, F1 score, and precision. Based on experimental results, the suggested method is 99.5% more accurate than pigeon-inspired optimization (PIO), Res-TranBiLSTM, and blockchain-based African buffalo (BbAB), which are 85.4%, 92.5%, and 85%, respectively.
Volume: 14
Issue: 3
Page: 722-732
Publish at: 2025-09-01

AI-driven solutions for Li-ion battery performance and prediction

10.11591/ijape.v14.i3.pp569-578
Sthitprajna Mishra , Chinmoy Kumar Panigrahi , Subhra Debdas , Atri Bandyopadhyay , Srikanth Velpula , Amit Kumar Sahoo , Pabitra Kumar Tripathy
Batteries serve as crucial power sources for essential portable devices like electric vehicles, smartphones, and laptops. The widespread adoption of Li-ion batteries, while beneficial, has unfortunately led to a surge in adverse incidents. The sudden failure of batteries in both industrial and lightweight applications poses significant economic risks across various industries. Consequently, researchers are intensifying their focus on enhancing battery state estimation, management systems, and predicting remaining useful life (RUL). This paper is structured into three main sections. Firstly, it delves into the acquisition of battery data, encompassing both commercially available and freely accessible Li-ion battery datasets. Secondly, the exploration extends to techniques for estimating battery states through advanced battery management systems. The paper investigates battery RUL estimation, categorizing and evaluating diverse prognostic methods applied to Li-ion batteries based on crucial performance parameters. The review includes scrutiny of commercially and publicly available datasets for various battery models and conditions, considering different battery states and the role of advanced battery management system (BMS). In the final section, the paper concludes with a comparative analysis of Li-ion battery RUL prediction, incorporating exploration into various RUL prediction algorithms, and mathematical models, and introducing an AI-based cloud monitoring system.
Volume: 14
Issue: 3
Page: 569-578
Publish at: 2025-09-01

Single photovoltaic panel constant regulated voltage based on modified DC-DC buck-boost converter topology

10.11591/ijape.v14.i3.pp620-630
Ro’ad Baladi Al Komar , Arwindra Rizqiawan
This research proposes a single photovoltaic panel constant regulated voltage based on novel topology. A modified DC-DC buck-boost converter was chosen because characteristics of voltage boost and low input current ripple. A comprehensive analysis of the proposed converter cells was elaborated in this study. Furthermore, a control technique is designed for the proposed converter. A double-loop control method using proportional integral (PI) is employed in this research. The outer loop controls the output voltage, while the inner loop is used to control the inductor current. By employing double-loop control, the presence of ripple current and voltage can be reduced even further. Simulation and experimental results validate the converter’s effectiveness, demonstrating stable voltage output under varying input voltage (33-36 V) and load conditions, maintaining a 40 V output with an overshoot within ±5%. The results show that the modified buck-boost converter can achieve improved efficiency and ripple reduction compared to conventional models, making it a viable solution for renewable energy systems.
Volume: 14
Issue: 3
Page: 620-630
Publish at: 2025-09-01
1 2 3
...
2027 2028
Show 166 of 2028

Discover Our Library

Embark on a journey through our expansive collection of articles and let curiosity lead your path to innovation.

Explore Now
Library 3D Ilustration