Claim Missing Document
Check
Articles

Generative artificial intelligence personalization and consumer behavior analytics Endang Tjahjaningsih; Elen Puspitasari; Felix Andreas Sutanto; Dewi Handayani Untari Ningsih; Dwi Budi Santoso
IAES International Journal of Artificial Intelligence (IJ-AI) Vol 15, No 5: October 2026
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijai.v15.i5.pp4919-4932

Abstract

Batik micro, small and medium enterprises (MSMEs) sustain Indonesia's textile heritage and local livelihoods but struggle with customer retention due to conventional marketing and limited artificial intelligence (AI) adoption. The shift to digital commerce requires affordable, culturally relevant AI strategies. This study proposes an integrated framework combining recency, frequency, monetary (RFM) analytics, K-means clustering (K = 5), random forest churn prediction, and transformer-based generative AI for personalized promotions delivered via WhatsApp business API. Evaluated over 12 months with 20 batik MSMEs in Central Java (300 customers, 873 transactions), the framework employed temporal data splitting (9 months training, 3 months testing) and A/B testing against generic campaigns. Human evaluation assessed message relevance and cultural appropriateness, while all data were anonymized with informed consent. Results show improvements of +17% in retention prediction accuracy, +23% in click-through rate, and +21% in purchase frequency versus baselines (p < 0.05, Cohen's d = 0.68, power > 0.8). The system costs below USD 20 per MSME per month, offering a scalable, human-centered pathway to strengthen loyalty in emerging markets.
Comparative Evaluation of TensorFlow Lite Quantization Variants for MobileNetV2-Based Batik Motif Classification: Accuracy, Model Size, and Inference Latency Dewi Handayani Untari Ningsih; Eri Zuliarso; Jati Sasongko Wibowo; Dwi Budi Santoso; Hari Murti
JURIKOM (Jurnal Riset Komputer) Vol. 13 No. 5 (2026): Oktober 2026
Publisher : Universitas Budi Darma

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30865/jurikom.v15i5.10047

Abstract

Convolutional Neural Network (CNN) models achieve high accuracy in visual recognition, but their size and computational cost limit deployment on mobile devices, so batik motif recognition remains difficult to bring to artisans and small creative enterprises. This study evaluates how post-training quantization with TensorFlow Lite affects the classification accuracy, model size, and inference latency of a MobileNetV2 transfer-learning model for batik motif recognition, in order to identify the deployment configuration that offers the best balance among the three. The experimental workflow adapted a pre-trained MobileNetV2 network to the Batik Wastra Nusantara dataset, which covers batik designs originating from different parts of Indonesia with substantial variation in color, ornamentation, and pattern complexity, converted the trained model into Dynamic Range, Float16, and Int8 TensorFlow Lite variants, and compared them under identical conditions. Evaluation on 644 valid images from 14 motif classes produced a test accuracy of 43.08%, corresponding to 28 of the 65 test images, a macro F1-score of 41.65%, and a weighted F1-score of 40.39%. Across the four model formats the accuracy spans only three test images, from 41.54% to 46.15%, a range that lies well inside the confidence interval implied by a 65-image test subset, so the variants cannot be separated on accuracy at this sample size. The deployment metrics differ substantially: relative to the unquantized reference, the Float16 variant reduced model size from 19.04 MB to 4.29 MB (77.47%) and inference latency from 205.52 ms to 28.60 ms (86.08%), while the Int8 variant was the fastest at 14.87 ms and the Dynamic Range variant the smallest at 2.41 MB. These findings indicate that the benefit of TensorFlow Lite optimization lies in storage and latency rather than in accuracy, and that the accuracy attainable on this dataset is constrained by its limited size and the visual similarity among motifs. The study contributes the deployment measurements, namely model size and inference latency for each quantization variant, that comparable batik classification studies do not report, and identifies dataset expansion as the prerequisite for reaching production-level accuracy in AI-based digital innovation for Indonesia’s creative industries.
Analisis Throughput Distribusi Jaringan Nirkabel Pada Politeknik Bumi Akpelni M. Rifki Wardana; Dwi Budi Santoso
Jurasik (Jurnal Riset Sistem Informasi dan Teknik Informatika) Vol 8, No 2 (2023): Edisi Agustus
Publisher : STIKOM Tunas Bangsa Pematangsiantar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30645/jurasik.v8i2.640

Abstract

Wireless networks, also known as wireless or Wi-Fi networks, are networks where users of electronic devices can connect and communicate with each other without using physical cables or wired connections. One of the factors that supports the capabilities of wireless networks is throughput. In the analysis of wireless network throughput, researchers conducted a study at Politeknik Bumi Akpelni, a maritime college. The research was conducted using computer simulations or field testing to examine factors that affect throughput, such as bandwidth, latency, and user density. The testing was performed by evaluating the wireless network transmitted by an access point device with the IEEE 802.11 ac standard, which operates at two frequencies: 2.4 GHz and 5 GHz. The testing was divided into three stages: testing wireless network throughput with obstacles (walls), testing wireless network throughput without obstacles, and testing wireless network throughput based on user density. The test results were categorized as very good, but it should be noted that the amount of bandwidth affected the quality of the throughput obtained. The more users there are, the smaller the throughput obtained.