Claim Missing Document
Check
Articles

Found 5 Documents
Search

PERBANDINGAN METODE PCA-SVM DAN SVM UNTUK KLASIFIKASI INDEKS KEPUASAN MASYARAKAT TERHADAP LAYANAN PENDIDIKAN DI KABUPATEN JENEPONTO Nur Ikhwana; Muhammad Nusrang; Sudarmin Sudarmin
VARIANSI: Journal of Statistics and Its application on Teaching and Research Vol 3, No 3 (2021)
Publisher : Universitas Negeri Makassar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35580/variansiunm22988

Abstract

Support Vector Machine (SVM) is one of the classification methods used to find the best hyperplane by maximizing the distance between classes. SVM aims to build a model that can predict the given test data. The SVM method can be implemented easily and the testing time is short, but it needs to reduce the computation burden. One way that can be done is to perform feature extraction to get the main characteristics of the data. The method that can be used to extract features is Principal Component Analysis (PCA). PCA is used to reduce the dimensions of data which are generally used in numerical scale data. If the data in the study used categorical data, then the PCA used was Nonlinear PCA. The data used in this study is the Community Satisfaction Survey data in Jeneponto Regency. This study compares the PCA-SVM and SVM methods for the classification of the Jeneponto Regency Community Satisfaction Index. The overall PCA-SVM classification results are better than SVM with 100% accuracy.
Trend, Cycle, and Forecasting Analysis of Monthly Inflation in Indonesia Using the Hodrick–Prescott Filter and ARIMA Nur Ikhwana; Annisa Syalsabila; Nalto Batty Mangiri; Lalu Ramzy Rahmanda
VARIANSI: Journal of Statistics and Its application on Teaching and Research Vol. 8 No. 01 (2026)
Publisher : Program Studi Statistika Fakultas MIPA UNM

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35580/variansiunm526

Abstract

This study aims to analyze the structure of inflation and forecast monthly inflation in Indonesia using a time series approach. The method used is the Hodrick–Prescott Filter to decompose data into trend and cycle components, and the ARIMA model to forecast inflation. The data used is monthly inflation data for the period 2010–2025. The decomposition results show that inflation has a relatively stable long-term trend with short-term fluctuations reflecting the presence of economic shocks. Based on model identification, the best model is ARIMA(2,0,1)(1,0,1)[12] which is able to capture past influences, seasonal components, and short-term shocks. The evaluation results show that the model meets the white noise assumption and is suitable for use in forecasting. The forecasting results show that inflation tends to be stable with a moderate increasing tendency, although uncertainty increases over longer periods. This study shows that the combination of structural analysis and time series modeling provides a more comprehensive understanding of inflation dynamics and produces relevant predictions to support decision making.
Forecasting Monthly Red Chili Prices in South Sulawesi Using Prophet Model with Time Series Cross-Validation Nur Ikhwana; Agung Triutomo; Annisa Syalsabila; Lalu Ramzy Rahmanda
ARRUS Journal of Mathematics and Applied Science Vol. 6 No. 1 (2026)
Publisher : PT ARRUS Intelektual Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35877/mathscience4893

Abstract

This study examines monthly red chili price movements in South Sulawesi using the Prophet forecasting model. Daily price data from the National Strategic Food Price Information Center (PIHPS) covering January 2020 to May 2026 were aggregated into 77 monthly observations. Missing values were handled using linear interpolation and Last Observation Carried Forward (LOCF) before modeling. The Prophet model using for forecasting and time series cross-validation was used as a validation method. The model performance was evaluated by Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE). The results indicate that the model produced an average RMSE = 11,024.68 IDR/kg, MAE = 9,139.04 IDR/kg and MAPE = 23.92%, suggests an acceptable forecasting performance for a highly volatile agricultural commodity. Results pattern demonstrate an annual seasonality on red chili prices, usually lowest price occurs in September and highest one in March of next year. The 12 months projections also foresee a maximum price of 72,597 IDR/kg in March 2027. These results show that the Prophet model can capture trend and seasonality, especially in predicting red chili prices of South Sulawesi.
Climate Risk in ESG Portfolios: Evidence from Indonesian Equity Markets Annisa Syalsabila; Nur Ikhwana; Lalu Ramzy Rahmanda
CAUCHY: Jurnal Matematika Murni dan Aplikasi Vol 11, No 2 (2026): CAUCHY: JURNAL MATEMATIKA MURNI DAN APLIKASI
Publisher : Mathematics Department, Maulana Malik Ibrahim State Islamic University of Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.18860/cauchy.v11i2.45016

Abstract

We examine climate variables in a pipeline for dynamic Value-at-Risk (VaR) of the SRI-KEHATI ESG index and the IHSG (Jakarta Composite Index). Using 1,499 daily observations (20142025), we combine the Least Absolute Shrinkage and Selection Operator (LASSO) over 50 candidate predictors, a quantile regression neural network (QRNN) for VaR at three quantile levels, and SHAP for attribution. LASSO removes all climate variables from both indices. For IHSG, the partial correlation of clean-energy returns falls from 0.121 to 0.0005 once domestic equity factors enter, consistent with domestic absorption, though constituent overlap cannot exclude a purely compositional explanation. A diagnostic quantile regression shows climate contributions rising 2.82-fold from VaR 90% to VaR 99%, but the pattern is flat under the QRNN itself (ratio 1.03). In the main backtest, QRNN has lower pinball loss than Historical Simulation and GARCH(1,1)-t in all 12 DieboldMariano comparisons (p 0.001), although it uses same-day predictors the benchmarks lack and the advantage was not reproduced in the ablation. In a common-sample ablation (n = 249 of 300 nominal test days), LASSO pre-selection is associated with 4 of 6 Kupiec passes versus 0 of 6 without it, but IHSG still fails at two of three quantile levels. Whether SRI-KEHATI screening isolates climate risk was not tested; mediation covers IHSG only.
Perbandingan Model Value-at-Risk (VaR) Hybrid GARCH-EVT dan Model Standar dalam Pengukuran Risiko Ekstrem pada Portofolio Saham Sektoral di Indonesia Annisa Syalsabila; Nur Ikhwana; Agung Tri Utomo; Lalu Ramzy Rahmanda; Zulkifli Rais
VARIANSI: Journal of Statistics and Its application on Teaching and Research Vol. 7 No. 03 (2025)
Publisher : Program Studi Statistika Fakultas MIPA UNM

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35580/variansiunm461

Abstract

This study aims to construct an optimal portfolio and compare the accuracy of various Value-at-Risk (VaR) models in measuring the risk of stock portfolios in the Indonesia Stock Exchange (IDX). The optimal portfolio is formed using the Minimum Variance Portfolio (MVP) method based on 11 sector-representative stocks for the period 2019–2025. The risk performance of this portfolio is then evaluated using six VaR models: Variance–Covariance (VC), Historical Simulation (HS), Monte Carlo (MC), GARCH (1,1), Extreme Value Theory (EVT-GPD), and the hybrid GARCH–EVT model. Model accuracy is assessed through backtesting using the Kupiec Proportion of Failures (POF) test and the Christoffersen Conditional Coverage (CC) test at the 95% and 99% confidence levels. The optimization results indicate that the MVP portfolio is dominated by defensive sectors such as consumer non-cyclicals (ICBP.JK) and large-cap banking (BBCA.JK). Backtesting results show that although all models perform adequately at the 95% level, standard models (VC, MC, GARCH) fail to capture extreme risk at the 99% level. In contrast, the GARCH–EVT model satisfies the backtesting criteria and emerges as the most accurate and superior model for predicting extreme losses.Penelitian ini bertujuan untuk membangun portofolio optimal dan membandingkan akurasi berbagai model Value-at-Risk (VaR) dalam mengukur risiko portofolio saham di Bursa Efek Indonesia (BEI). Portofolio optimal dibentuk menggunakan metode Minimum Variance Portfolio (MVP) dari 11 saham perwakilan sektor periode 2019-2025. Kinerja risiko portofolio ini kemudian diukur menggunakan enam model VaR: Variance-Covariance (VC), Historical Simulation (HS), Monte Carlo (MC), GARCH (1,1), Extreme Value Theory (EVT-GPD), dan model hybrid GARCH-EVT. Akurasi model diuji menggunakan backtesting Uji Kupiec (POF) dan Uji Christoffersen (CC) pada tingkat kepercayaan 95% dan 99%. Hasil optimisasi menunjukkan portofolio MVP didominasi oleh sektor defensif seperti consumer non-cyclicals (ICBP.JK) dan perbankan big-cap (BBCA.JK). Hasil backtesting menunjukkan bahwa meskipun semua model akurat pada tingkat 95%, model standar (VC, MC, GARCH) gagal mengukur risiko ekstrem pada tingkat 99%. Sebaliknya, model GARCH-EVT terbukti memenuhi uji dan menjadi model yang paling akurat dan superior untuk memprediksi kerugian ekstrem.