Article Per Year (5 Year)
p-Index From 2021 - 2026
1.633
P-Index
This Author published in this journals
All Journal JSI: Jurnal Sistem Informasi (E-Journal) Jurnal Geocelebes Martabe : Jurnal Pengabdian Kepada Masyarakat Jambura Journal of Mathematics Jurnal Ilmiah Matematika Enthusiastic : International Journal of Applied Statistics and Data Science Malcom: Indonesian Journal of Machine Learning and Computer Science Multi Proximity: Jurnal Statistika Universitas Jambi Mathematical Journal of Modelling and Forecasting Journal of Statistics and Data Science
Lestari, Wina Ayu
Universitas Bengkullu
Agriculture, Biological Sciences & Forestry Arts Humanities Computer Science & IT Decision Sciences, Operations Research & Management Earth & Planetary Sciences Economics, Econometrics & Finance Education Engineering Library & Information Science Mathematics Social Sciences Other

Published : 10 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 10 Documents
Search

 1 
Forecasting Export Value of Bengkulu Province Through Pulau Baai Harbour with ARIMA, ANN, and Hybrid ARIMA-ANN Approach Lestari, Wina Ayu; Nugroho, Sigit; Widodo, Fanani Haryo Widodo
Journal of Statistics and Data Science Vol. 3 No. 1 (2024)
Publisher : UNIB Press

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33369/jsds.v3i1.41289

Abstract

Forecasting is a process of predicting future events based on past event data. One of the time series models that can be used for forecasting is the Autoregressive Integrated Moving Average (ARIMA). The advantages of ARIMA are in the accuracy and flexibility of its forecasting in representing several different types of time series, but the main limitation is the linear form of the model which causes ARIMA to be unable to capture non-linear patterns in the data. An alternative model for time series modeling is Artificial Neuron Network (ANN). ANN can overcome the weaknesses of ARIMA, but cannot handle linear and nonlinear patterns of the data simultaneously. As an effort to improve forecasting accuracy, Hybrid ARIMA-ANN is carried out by taking advantage of the supremacy of ARIMA and ANN. This study aims to obtain the best model for forecasting the export value of Bengkulu Province, a model generated by the time series data of export values issued by Pulau Baai Harbour from January 2014 to June 2022. The result shows that the best model for predicting the export value of Bengkulu Province is the ARIMA-ANN hybrid model with MAAPE of 0.5289 and MASE of 0.7664.
Microclimatic Temperature Variability and Trends in Bengkulu Province: ANOVA and Regression-Based Analysis Norfahmi, Siti Hairunnisa; Samdara, Rida; Supiyati, Supiyati; Lestari, Wina Ayu
Jambura Journal of Mathematics Vol 7, No 2: August 2025
Publisher : Department of Mathematics, Universitas Negeri Gorontalo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37905/jjom.v7i2.33376

Abstract

This study investigates the microclimatic variability and trends of air temperature across three meteorological stations—Fatmawati, Bengkulu, and Kepahiang—in Bengkulu Province, Indonesia. Using five years of daily data (June 2020 to May 2025), minimum (Tmin), maximum (Tmax), and average (Tavg) temperatures were analyzed to understand both spatial patterns and temporal changes in surface air temperature. One-way ANOVA was conducted to assess whether mean temperatures differed significantly across stations, followed by Tukey  post hoc test for pairwise comparisons. The analysis revealed a consistent and statistically significant difference in all temperature variables (p 0.05), particularly between the inland highland station (Kepahiang) and the two coastal stations. In addition, monthly averages of Tavg were analyzed using simple linear regression, with significance tested via regression-based ANOVA. All three stations exhibited statistically significant warming trends (p 0.005), with slopes ranging from +0.0152 to +0.0213 °C/month (~0.18–0.26 °C/year), despite relatively modest coefficients of determination (R² = 0.14–0.24). These results highlight a dual climatic dynamic in the region: strong seasonal and spatial variability, overlaid with emerging baseline warming. The study underscores the importance of localized climate analysis for adaptation planning, particularly in topographically diverse tropical regions facing increased exposure to climate variability and change.
Comparative Analysis of SARIMA, FFNN, and Hybrid Models for Sea Surface Temperature Prediction at Enggano Island (2018–2024) Natisharevi, Raditya Janaloka; Rizal, Jose; Firdaus, Firdaus; Novianti, Pepi; Lestari, Wina Ayu
JURNAL GEOCELEBES Vol. 9 No. 2: October 2025
Publisher : Departemen Geofisika, FMIPA - Universitas Hasanuddin, Makassar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70561/geocelebes.v9i2.46445

Abstract

Sea Surface Temperature (SST) is a key oceanographic variable that influences fish distribution and the livelihoods of coastal communities. On Enggano Island, where most residents rely on fishing, SST is critical for identifying optimal fishing grounds due to limited accessibility and high operational costs. Accurate modeling and forecasting of SST are therefore essential for effective fisheries management and sustainable resource use. This study analyzes and predicts monthly SST patterns in Enggano Island using Seasonal Autoregressive Integrated Moving Average (SARIMA), Feed Forward Neural Network (FFNN), and Hybrid SARIMA-FFNN models. SARIMA effectively captures linear trends and seasonal variations but struggles with nonlinear dynamics and requires statistical assumptions. Conversely, FFNN models nonlinear relationships without such assumptions but is less efficient in representing linear and seasonal structures. The hybrid SARIMA-FFNN combines the strengths of both approaches, integrating linear-seasonal accuracy with nonlinear adaptability. Monthly SST data from January 2018 to December 2024, covering northern, eastern, southern, and western regions of Enggano Island, were analyzed. Results show that all models achieved high predictive accuracy, with MAPE values below 10%. Based on RMSE, FFNN outperformed the other models across all regions (north: 1.173, east: 0.999, south: 1.245, west: 1.049), confirming FFNN as the most accurate model for SST prediction. Predicted SST values across the four regions exhibited only minor differences, offering fishermen flexibility in selecting fishing grounds. Sustainable fishing strategies should also consider species-specific temperature preferences and other ecological factors influencing fish distribution.
DESA CINTA STATISTIK: INFOGRAFIS DATA STATISTIK DESA PANCA MUKTI DENGAN CANVA Ayu Lestari, Wina; Sunandi, Etis; Agwil, Winalia; Hidayati, Nurul; Firdaus, Firdaus; Fairuzinda, Athaya; Pandu Winata, Aji; Yulian Utami, Rahmi
Martabe : Jurnal Pengabdian Kepada Masyarakat Vol 8, No 6 (2025): MARTABE : JURNAL PENGABDIAN KEPADA MASYARAKAT
Publisher : Universitas Muhammadiyah Tapanuli Selatan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31604/jpm.v8i6.2661-2666

Abstract

Kegiatan Pengabdian kepada Masyarakat ini bertujuan untuk meningkatkan kemampuan aparatur desa dalam mengolah, menganalisa dan menyajikan data statistik sektoral dalam bentuk infografis desa. Perangkat desa belum memiliki kemampuan melakukan visualisasi data statistik desa dengan infografis menjadi latar belakang dilaksanakannya pelatihan ini. Metode pelaksanaan kegiatan ini meliputi persiapan, pelaksanaan, dan evaluasi. Hasil pretest dan posttest digunakan untuk mengevaluasi efektivitas pelatihan serta mengukur sejauh mana peningkatan pemahaman dan keterampilan peserta setelah mengikuti kegiatan pengabdian kepada masyarakat. Berdasarkan uji Wilcoxon Signed-Rank diketahui bahwa kegiatan pengabdian kepada masyarakat memberikan pengaruh yang signifikan terhadap pemahaman dan keterampilan peserta tentang infografis.
Application of Support Vector Regression (SVR) with a Radial Basis Function (RBF) Kernel for Predicting the Global Happiness Index Aji Pandu Winata; Nala Kamila Azizy; Wina Ayu Lestari; Riwi Dyah Pangesti
Mathematical Journal of Modelling and Forecasting Vol. 4 No. 1 (2026): June 2026
Publisher : Universitas Negeri Padang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24036/mjmf.v4i1.54

Abstract

The Global Happiness Index is widely used to measure countries' well-being across social, economic, and health-related dimensions. The complex and non-linear relationships among these dimensions often limit the predictive performance of conventional linear regression models. This study aims to evaluate the effectiveness of Support Vector Regression (SVR) with a Radial Basis Function (RBF) kernel in predicting Global Happiness Index scores. The study used data from 158 countries obtained from Kaggle, including GDP per capita, social support, healthy life expectancy, freedom, trust in government, and generosity as predictor variables. Data preprocessing was performed before splitting the dataset into training and testing sets, and the optimal SVR parameters were determined using Grid Search with K-fold cross-validation. The optimal SVR-RBF model produced an RMSE of 0.4462 and an MAE of 0.3829 on the testing data. In addition, the model achieved an R² value of 0.8328, indicating that it explained 83.28% of the variation in Global Happiness Index scores. These results suggest that SVR with an RBF kernel is an effective approach for modeling complex nonlinear relationships and can be used as a reliable tool for predicting national happiness levels.
Analisis Faktor-Faktor yang Mempengaruhi Ketahanan Pangan Kabupaten/Kota di Indonesia Menggunakan Algoritma Classification and Regression Tree (CART) Wina Ayu Lestari; Aulia Rahma Dewi; Dinda Aprilia; Riwi Dyah Pangesti
Jurnal Ilmiah Matematika Vol. 13 No. 1 (2026)
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26555/jim.v13i1.32082

Abstract

Ketahanan pangan merupakan isu strategis dalam pembangunan nasional karena berkaitan dengan kesejahteraan dan stabilitas sosial ekonomi, namun masih terdapat disparitas antar kabupaten/kota di Indonesia. Penelitian ini bertujuan menganalisis faktor-faktor yang memengaruhi ketahanan pangan menggunakan algoritma Classification and Regression Tree (CART). Data bersumber dari publikasi resmi Badan Pusat Statistik yang mencakup 512 kabupaten/kota, dengan variabel respon berupa komposit ketahanan pangan dan variabel prediktor meliputi kedalaman kemiskinan, persentase pengeluaran pangan, serta persentase stunting. Analisis dilakukan melalui tahapan preprocessing, penyederhanaan kategori ketahanan pangan untuk mengatasi ketidakseimbangan kelas, pembagian data pelatihan dan pengujian, pembangunan serta pemangkasan pohon CART, dan evaluasi menggunakan confusion matrix. Hasil menunjukkan model terbaik memiliki akurasi 84,47% dan macro F1 sebesar 0,5749. Kedalaman kemiskinan menjadi faktor paling dominan, diikuti stunting dan pengeluaran pangan. Wilayah dengan kemiskinan tinggi cenderung berada pada kategori ketahanan pangan rentan, sedangkan wilayah dengan kemiskinan rendah dan kondisi gizi lebih baik cenderung berada pada kategori ketahanan pangan tahan. Dengan demikian, CART terbukti efektif dan interpretatif dalam mengidentifikasi faktor utama ketahanan pangan sebagai dasar kebijakan berbasis data.
Implementation of Multivariate Small Area Estimation on Socio-Economic Indicators Firdaus, Firdaus; Nugroho, Sigit; Ayu Lestari, Wina; Pandu Winata, Aji
Enthusiastic : International Journal of Applied Statistics and Data Science Volume 6 Issue 1, April 2026
Publisher : Universitas Islam Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.20885/enthusiastic.vol6.iss1.art9

Abstract

The welfare of a region’s population is generally measured through socio-economic indicators such as per capita expenditure, access to education, and access to safe drinking water. This study aimed to estimate socio-economic indicators in Bengkulu Province at the small-area level using the multivariate small area estimation (SAE) approach under the empirical best linear unbiased prediction (EBLUP) framework. The analysis was based on secondary data obtained from BPS-Statistics Indonesia, covering the 2024 observation period, with indicators aggregated at the district level. Area-level auxiliary variables derived from the Village Potential Statistics (Potensi Desa, PODES) were incorporated to improve estimation accuracy by borrowing strength across correlated indicators and areas. Results showed that indirect estimation substantially improved the precision of direct estimates. The mean squared error (MSE) for the school participation rate decreased from 0.260 to 0.040, while the MSE for access to safe drinking water decreased from 0.00550 to 0.00098, indicating efficiency gains of approximately 85% for both indicators. Overall, the multivariate EBLUP approach yielded more stable and reliable estimates than direct estimation. Therefore, this approach is recommended for generating small area estimates of correlated welfare indicators when sample sizes are limited, thereby providing more accurate information for regional planning and policy formulation.
Perbandingan Model Naïve, Model Linier Umum dan Model Linier Efek Random pada Data Longitudinal Saham Bulanan Perusahaan Manufaktur Indonesia Redha, Abdul Rahim; Mezilano, M. Zidan; Fhatiha, Shyalaisha Amani; Afriasih, Zakila; Lestari, Wina Ayu
Multi Proximity: Jurnal Statistika Vol. 4 No. 1 (2025): Multiproximity
Publisher : Universitas Jambi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.22437/6qgwxv70

Abstract

Penelitian ini bertujuan untuk membandingkan kinerja Model Naïve, Model Linier Umum dan Model Linier Efek Random dalam menganalisis data longitudinal harga saham bulanan perusahaan manufaktur Indonesia periode 2020–2025. Data yang digunakan merupakan data sekunder dari tiga perusahaan yang terdaftar di Bursa Efek Indonesia, dengan variabel respon berupa Adjusted Close dan lima variabel prediktor yaitu Open, High, Low, Close, dan Volume. Analisis diawali dengan eksplorasi data menggunakan profile plot dan LOWESS plot untuk melihat pola tren dan fluktuasi. Selanjutnya dibangun Model Naïve menggunakan metode OLS, Model Linier Umum dengan berbagai struktur korelasi (Compound Symmetry, Autoregressive (AR ), dan Eksponensial), serta Model Linier Efek Random dengan dan tanpa struktur korelasi menggunakan metode REML. Hasil penelitian menunjukkan bahwa variabel High berpengaruh negatif dan signifikan, sedangkan Close berpengaruh positif dan signifikan terhadap harga saham pada seluruh model. Berdasarkan kriteria Akaike Information Criterion (AIC), Model Linier Efek Random dengan struktur korelasi Autoregressive (AR ) dan Eksponensial memberikan nilai AIC terkecil, sehingga menjadi model terbaik. Temuan ini menegaskan pentingnya mempertimbangkan heterogenitas antar perusahaan dan korelasi temporal dalam analisis Data Longitudinal sektor keuangan
Evaluasi Model Regresi Linier Berganda dalam Memprediksi Performance Index Menggunakan Cross Validation: Evaluation of Multiple Linear Regression Models in Predicting the Performance Index Using Cross-Validation Darani, Viche; Safebsi, Chika Mutia; Karimah, Dina; Putri, Deswita Dwi; Sunandi, Etis; Lestari, Wina Ayu
MALCOM: Indonesian Journal of Machine Learning and Computer Science Vol. 6 No. 4 (2026): MALCOM October 2026
Publisher : Institut Riset dan Publikasi Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.57152/malcom.v6i4.2802

Abstract

Penelitian ini bertujuan untuk membangun dan mengevaluasi model regresi linier berganda dalam memprediksi Performance Index berdasarkan lima variabel independen, yaitu Study Hours, Extra Participation, Papers Practiced, Previous Marks, dan Sleeping Hours. Data yang digunakan bersumber dari Kaggle dan terdiri dari 1500 observasi yang telah melalui tahap preprocessing sehingga seluruh data bersih dan siap digunakan dalam analisis. Pengujian asumsi klasik yang mencakup uji linearitas, normalitas residual, multikolinearitas, heteroskedastisitas, dan autokorelasi menunjukkan bahwa seluruh asumsi terpenuhi. Berdasarkan hasil penelitian diperoleh nilai R-squared sebesar 0.9888, yang berarti 98.88% variasi pada variabel dependen dapat dijelaskan oleh kelima variabel independen, dengan Previous Marks sebagai variabel independen paling dominan. Seluruh variabel independen berpengaruh signifikan secara parsial maupun simultan terhadap variabel dependen. Evaluasi performa model menghasilkan nilai MAE sebesar 1.6081, MSE sebesar 4.1007, dan RMSE sebesar 2.0250. Validasi model menggunakan 10-fold cross-validation menghasilkan nilai metrik yang sangat mendekati hasil evaluasi awal, membuktikan bahwa model bersifat stabil dan tidak mengalami overfitting. Kebaruan penelitian ini terletak pada penerapan 10-fold cross-validation secara eksplisit dengan pelaporan metrik per fold serta perbandingan kuantitatif antara hasil evaluasi awal dan hasil cross-validation sebagai bukti ketiadaan overfitting.
Implementasi Artificial Neural Network dengan menggunakan Deep Multi-Layer Perceptron untuk Estimasi Konsumsi Energi pada Industri Manufaktur Baja Athaya Fairuzindah; Risfa Fadila; Aurel Giovani Aria; Wina Ayu Lestari; Riwi Dyah Pangesti
JSI: Jurnal Sistem Informasi (E-Journal) Vol 18 No 1 (2026): JSI: Jurnal Sistem Informasi (E-Journal)
Publisher : Jurusan Sistem Informasi Fakultas Ilmu Komputer Universitas Sriwijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.18495/jsi.v18i1.460

Abstract

Industri baja memegang peranan strategis dalam menunjang pertumbuhan ekonomi dan pembangunan infrastruktur nasional. Dalam proses pengolahan baja bergantung pada pengoperasian mesin dan peralatan yang membutuhkan konsumsi energi dalam jumlah besar. Tingginya konsumsi energi tersebut memerlukan upaya pengelolaan dan efisiensi energi untuk meningkatkan kinerja industri manufaktur baja. Salah satu pendekatan yang dapat digunakan untuk estimasi konsumsi energi pada industri manufaktur baja adalah Artificial Neural Network dengan Deep Multi-Layer Perceptron, yang dapat merepresentasikan hubungan nonlinier yang sangat kompleks pada industri baja. Penelitian ini mengusulkan implementasi Artificial Neural Network untuk konsumsi energi pada industri manufaktur baja dari perusahaan DAEWOO Steel Co. Ltd. Sehingga, diperoleh nilai RMSE sebesar 0.9261, nilai MAPE sebesar 63.46%., dan R-Squared sebesar 0.9992. Dapat disimpulkan bahwa penggunaan energi di perusahaan baja 99.92% dipengaruhi oleh Lagging Current Reactive Power, Leading Current Reactive Power, CO2, Lagging Current Power Factor, Leading Current Power Factor, NSM, Week, Day of Week, dan Load Type.
Co-Authors Afriasih, Zakila Agwil, Winalia Aji Pandu Winata Athaya Fairuzindah Aulia Rahma Dewi Aurel Giovani Aria Darani, Viche Dinda Aprilia Fairuzinda, Athaya Fhatiha, Shyalaisha Amani Firdaus Firdaus Jose Rizal Karimah, Dina Mezilano, M. Zidan Nala Kamila Azizy Natisharevi, Raditya Janaloka Norfahmi, Siti Hairunnisa Nurul Hidayati Pandu Winata, Aji Pepi Novianti Putri, Deswita Dwi Redha, Abdul Rahim Rida Samdara Risfa Fadila Riwi Dyah Pangesti Riwi Dyah Pangesti Safebsi, Chika Mutia Sigit Nugroho Sunandi, Etis Supiyati . Widodo, Fanani Haryo Widodo Yulian Utami, Rahmi