Jurnal Aplikasi Sains Data

Comparison of ARIMA and SARIMA Methods for Non-Oil and Gas Export Forecasting in East Java

2025-05-21T06:18:32+00:00

Forecasting plays a pivotal role in economic planning, particularly in aligning supply with demand and informing production decisions. This study aims to compare the performance of the Autoregressive Integrated Moving Average (ARIMA) and Seasonal ARIMA (SARIMA) models in forecasting the non-oil and gas export values of East Java, a region known for its dynamic trade activity. Using monthly time series data spanning from January 2007 to January 2024, sourced from the Central Statistics Agency (BPS) of East Java Province, this research conducts an in-depth analysis of forecasting accuracy and model suitability. Before model implementation, the dataset underwent several preprocessing steps to ensure its quality, including the handling of missing values and outlier adjustments. Both ARIMA and SARIMA models were developed, calibrated, and evaluated using standard forecasting performance metrics, namely Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). The ARIMA model exhibited consistently lower error rates across all three metrics, indicating its robustness in capturing the underlying patterns within the export data. In contrast, while the SARIMA model incorporated seasonal components, its performance did not surpass that of ARIMA in this specific case. The comparative findings suggest that, despite the seasonal nature of trade, the ARIMA model is more suitable for short-term forecasting of East Java’s non-oil and gas exports. This research contributes to the broader literature on economic forecasting by emphasizing the importance of selecting appropriate models based on data characteristics. Furthermore, the results provide valuable insights for policymakers and stakeholders engaged in export planning and regional trade development In this result the ARIMA model overcome the SARIMA with MAPE 0.116 to 0.983.

Implementation of Content-Based Filtering in Tourist Destination Recommendation System in Central Java

2025-05-23T15:03:21+00:00

Tourism is a potential sector that plays an important role in the regional economy with significant contributions to regional income and foreign exchange earnings. Central Java, as one of the provinces with great potential in the tourism sector, has a variety of tourist attractions that include natural, artificial, special interest destinations, and more. One effort to optimize the tourism sector in Central Java is to improve tourism information services by creating a recommendation system for tourist attractions in Central Java. This research aims to create a personalized recommendation system for tourist attractions in Central Java based on user preferences using content-based filtering methods and neural network machine learning. This method is used to analyze the features of tourist attractions and user preferences, and to generate relevant recommendations. The model is trained using Adam optimization with a learning rate of 0.01 and 300 epochs. The evaluation results show that this method can provide tourist attraction recommendations in Central Java that tend to match user preferences with relatively low error rates, as indicated by a Mean Squared Error (MSE) value of 0.1766. Thus, this research can contribute to optimizing the tourism sector in Central Java and guide individuals in finding tourist attractions that suit their individual preferences.

Comparative Analysis of Stochastic Gradient Descent Optimization and Adaptive Moment Estimation in Emotion Classification from Audio Using Convolutional Neural Network

2025-05-27T07:18:03+00:00

Emotion is a fundamental aspect of human life that profoundly shapes behavior, social interactions, and decision-making processes. The ability to effectively communicate and foster mutual understanding between individuals relies heavily on accurately recognizing and expressing emotions. Among various channels of emotional expression, sound stands out as a powerful and direct medium that reflects and conveys human emotional states. This makes audio-based emotion recognition a critical and rapidly evolving field of study. With the rapid advancements in information technology and artificial intelligence, research focused on recognizing emotions through sound signals has gained significant momentum. Machine learning algorithms, particularly deep learning models like neural networks, have demonstrated remarkable capabilities in identifying and classifying emotions expressed through multiple modalities such as text, images, videos, and especially audio signals. Within the family of neural networks, Convolutional Neural Networks (CNNs) have been especially effective for audio emotion classification, due to their strength in extracting hierarchical and spatial features directly from raw input data. This study specifically investigates the comparative effectiveness of two popular optimization algorithms—Stochastic Gradient Descent (SGD) and Adaptive Moment Estimation (Adam)—in training CNN models for emotion classification from audio recordings. Utilizing the Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS) dataset, experimental results indicate that CNNs trained with the SGD optimizer achieve an overall accuracy of 53%, surpassing the 48% accuracy achieved by Adam. These results underscore the potential advantages of SGD in fine-tuning deep learning models for audio-based emotion recognition. Consequently, researchers and practitioners are encouraged to consider SGD optimization to improve the performance and robustness of emotion classification systems based on audio data.

Application of Fuzzy Inference System for Quality Assessment of Formula Milk for Pregnant Women in Stunting Program

2025-05-27T07:57:22+00:00

Stunting remains a significant global public health challenge, affecting more than 149 million children under five years of age worldwide as reported by the United Nations in 2020. Indonesia alone accounts for approximately 6.3 million stunted children, highlighting the urgent need for effective intervention strategies. Stunting is primarily caused by chronic malnutrition during the first 1,000 days of life, which includes inadequate nutritional intake during pregnancy, poor infant feeding practices, and environmental factors such as inadequate sanitation. The 2022 Indonesian Nutrition Status Survey (SSGI) indicated a stunting prevalence of 21.6%, showing improvement from 24.4% in 2021, yet still significantly above the national target of 14% set for 2024. Given the critical role of maternal nutrition in reducing stunting risk, providing pregnant women with appropriate nutritional guidance is essential. This study aims to develop a decision support model using a Fuzzy Inference System (FIS) to assist pregnant women in selecting the most suitable formula milk based on nutritional value and affordability. The Mamdani FIS method was applied to analyze data from eight commercially available formula milk products. The evaluation measured the membership degrees corresponding to recommendation levels, factoring in both price and nutrition. The results identified Anmum Materna as the most favorable option, with a membership degree of 0.937, classified under the "Highly Recommended" category. This formula is priced at IDR 70,000 and contains a total nutritional value of 1024 grams, offering a balance of quality and affordability. This model demonstrates potential as a practical tool to support informed nutritional choices during pregnancy, contributing to stunting prevention efforts.

Application of K-Prototypes Clustermix Algorithm for Clustering Risk Factors of Diabetes Disease

2025-05-27T15:06:53+00:00

Diabetes mellitus (DM) is recognized as one of the most rapidly increasing chronic diseases worldwide, posing a significant public health challenge. According to the International Diabetes Federation (IDF), approximately 537 million people were living with diabetes mellitus globally, with projections estimating a rise to 643 million by 2030 and 783 million by 2045. Additionally, the World Health Organization (WHO) reported a 3% increase in mortality rates attributed to diabetes mellitus between 2000 and 2019, underscoring the urgent need for effective risk detection and management strategies. Early identification of risk factors is crucial to mitigating the impact of DM, and clustering analysis offers a promising method for stratifying patients based on risk profiles. This study employs the k-prototypes algorithm, which is particularly suited to clustering datasets with mixed numeric and categorical variables, to analyze DM risk factors. Utilizing data from the 2022 Behavioral Risk Factor Surveillance System (BRFSS) annual survey, the study examines a sample of 2,480 diabetes mellitus patients across the United States. The clustering analysis identified two optimal clusters (k=2) based on a high silhouette score of 0.821, indicating strong cluster cohesion and separation. Cluster 2, consisting of 77 patients, exhibited a higher risk profile for diabetes compared to Cluster 1, which included 2,403 patients. The clusters were characterized by significant differences in average values of key DM risk factors including weight, fruit and vegetable consumption, mental and physical health status, age, alcohol consumption, hypertension, smoking status, physical activity, mobility difficulties, sex, education level, income, and ethnicity. These findings highlight the utility of k-prototypes clustering in identifying high-risk DM subgroups to inform targeted prevention and intervention efforts.