Jurnal Galaksi

Retail Sales Forecasting Using ARIMA Based on Point-of-Sale Transaction Data

I Gede Sudiantara — 2026-08-31

Retail businesses generate large volumes of Point-of-Sale (POS) transaction data that can be utilized for sales forecasting and inventory planning. However, transaction-level data need to be transformed into an appropriate time-series structure before forecasting can be performed. This study aims to apply the Autoregressive Integrated Moving Average (ARIMA) method to historical POS transaction data for retail sales forecasting. The study used 4,443,905 transaction records from a retail store covering the period from 2018 to 2021. The research process adopted the Cross-Industry Standard Process for Data Mining (CRISP-DM), with the analysis focusing on data preparation, ARIMA modeling, and model evaluation. Sales data were aggregated into product-level time series, and the ARIMA model was identified using stationarity analysis, Autocorrelation Function (ACF), and Partial Autocorrelation Function (PACF). The resulting ARIMA(1,1,1) model was applied to five selected products represented by M01–M05. Model evaluation used the Ljung–Box test and Root Mean Square Error (RMSE), with forecasting conducted for January and February 2022. The results show that forecasting errors varied across products, indicating differences in forecasting performance according to their historical sales patterns. The study demonstrates the applicability of ARIMA for product-level retail sales forecasting based on historical POS transaction data.

Comparison of Grayscale and Binary Image Representations for Balinese Script Classification Using EfficientNetV2

I Made Dwi Putra Asana — 2026-08-31

Balinese script is a cultural heritage whose preservation increasingly depends on digital technology, yet automatic recognition of its characters remains difficult because many glyphs share similar strokes. This study compares two image representations, grayscale and binary, as input for classifying 28 classes of Balinese script consisting of eighteen basic characters (wreastra), six pengangge suara and four pengangge tengenan. A dataset of 1,337 images was resized to 224x224 pixels and augmented into 8,022 images, then classified using EfficientNetV2B0 with transfer learning, a batch size of 32, a learning rate of 0.001, a dropout rate of 0.3 and 30 epochs followed by fine tuning of the last thirty layers. Robustness was also examined through three hold-out splitting ratios of 80:10:10, 70:15:15 and 60:20:20, which produced accuracies of 90.55%, 89.55% and 88.81%. The grayscale representation reached a test accuracy of 91.92% with a macro F1-score of 92.11%, slightly higher than the binary representation with 91.67% and 91.78%. Deployment on new handwritten images revealed a domain gap that lowered recognition performance, indicating that data diversity is more decisive than the choice of representation.

Expert System Design for Laying Hen Disease Diagnosis Using Forward Chaining and Certainty Factor

Ketut Jaya Atmaja — 2026-08-31

Laying hens are a crucial part of the poultry industry, significantly contributing to the fulfillment of animal protein needs. However, the prevalence of various diseases remains a primary threat that causes a drastic decline in egg production and high mortality rates. The delay in diagnosis, exacerbated by the limited availability of veterinary specialists in rural areas, emphasizes the need for an accessible diagnostic tool. This research aims to develop a web-based expert system capable of diagnosing 11 types of laying hen diseases based on 27 observable symptoms. The system implements the Forward Chaining method as an inference engine to systematically trace disease rules starting from the user-inputted symptoms. Furthermore, the Certainty Factor (CF) method is integrated to calculate the confidence level of the diagnosis, utilizing the Measure of Belief (MB) and Measure of Disbelief (MD) values provided by an expert veterinarian. The system was tested using Blackbox testing for functional validation, which yielded 100% functionality. Furthermore, accuracy testing across 18 real-world test cases demonstrated that the system achieved an accuracy rate of 77.78%, with 14 cases perfectly matching the expert's diagnosis. This expert system proves to be highly effective in assisting farmers in performing early disease identification and taking timely preventive measures.