Claim Missing Document
Check
Articles

Found 24 Documents
Search

POVERTY PANEL DATA MODELING IN SOUTH SUMATERA Hidayati, Nurul; Fransiska, Herlin; Agwil, Winalia
BAREKENG: Jurnal Ilmu Matematika dan Terapan Vol 17 No 3 (2023): BAREKENG: Journal of Mathematics and Its Applications
Publisher : PATTIMURA UNIVERSITY

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30598/barekengvol17iss3pp1203-1214

Abstract

Poverty is still a major problem on Sumatra's island, despite abundant natural resources potential, such as mining and plantation products. Sumatra Island consists of 10 provinces divided into three regions: Northern Sumatra, Central Sumatra, and Southern Sumatra. The island of Sumatra has the highest number of poor people in the Southern Sumatra region, which reaches 2.88 million people. Poverty is an integrated concept with five dimensions: poverty, powerlessness, vulnerability to emergencies, dependency, and alienation, both geographically and sociologically. One method that can be used to analyze poverty data problems is panel data regression analysis, which combines two data, namely cross-sectional data and time series data. It is expected to produce more in-depth and comprehensive information, both the interrelationships between the variables and their development within a certain period. The panel data was related to poverty and included 60 districts in the southern Sumatra region from 2018 to 2020. This study aimed to model poverty panel data in Southern Sumatra. Three estimation methods were used in the panel data regression, including the Common Effect Model (CEM), Fixed Effect Model (FEM), and Random Effect Model (REM. The results of the model specification test show that the best model for estimating the percentage of poor people in the Southern Sumatra region is the Fixed Effect Model (FEM), with a value of R2 = 75.57%. The results of the significance test show that the variables that significantly influence the percentage of poverty in the Southern Sumatra region using the FEM model are the open unemployment rate (), life expectancy (), and the average length of schooling ().
MODELING CLUSTERWISE LINEAR REGRESSION ON POVERTY RATE IN INDONESIA Meylisah, Eni; Rini, Dyah Setyo; Fransiska, Herlin; Agwil, Winalia; Sartono, Bagus
BAREKENG: Jurnal Ilmu Matematika dan Terapan Vol 17 No 3 (2023): BAREKENG: Journal of Mathematics and Its Applications
Publisher : PATTIMURA UNIVERSITY

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30598/barekengvol17iss3pp1653-1662

Abstract

When a person's income is so low that it cannot cover even the most basic living expenses, they are said to be poor. Data on poverty levels and hypothesized causes are used in this study. If the data pattern forms clusters, one of the regression analyses that can be used is Clusterwise Linear Regression (CLR). Therefore, this study aimed to determine the poverty rate modeling in Indonesia with the CLR method. The results showed that the best model is with 3 clusters, that for cluster 1, the factors that significantly affect the percentage of poverty are the percentage of electricity users , the number of small and micro industries and the number of tourist villages n cluster 2, the amount of village tours . In cluster 3, the percentage of users of electricity and the percentage of villages that have mining and quarrying .
Analisis Perbandingan Metode K-Medoid dan Ward dalam Klasterisasi Nasabah Bank Churners Simanullang, Aditama Rouliber; Syah, Vivi Elvira Saputri; Afni, Della Nur; Popita, Resi; Sunandi, Etis; Agwil, Winalia
MATHunesa: Jurnal Ilmiah Matematika Vol. 13 No. 2 (2025)
Publisher : Universitas Negeri Surabaya

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

This study compares the K-Medoid and Ward clustering methods in segmenting bank churner customers. The dataset used is the Bank Churners dataset, consisting of 5.000 observations and six variables related to credit card usage. The results show that the Ward method has a higher silhouette coefficient compared to K-Medoid, making it more effective in forming homogeneous clusters. Ward clustering produces four distinct customer segments, which can help banks develop targeted retention strategies. Based on the analysis, promotional strategies and personalized services are recommended to reduce customer churn according to each cluster’s characteristics.
BIRESPONSE SPLINE TRUNCATED NONPARAMETRIC REGRESSION MODELING FOR LONGITUDINAL DATA ON MONTHLY STOCK PRICES OF THREE PRIVATE BANKS IN INDONESIA Pahlepi, Reza; Sriliana, Idhia; Agwil, Winalia; Oktarina, Cinta Rizki
BAREKENG: Jurnal Ilmu Matematika dan Terapan Vol 19 No 4 (2025): BAREKENG: Journal of Mathematics and Its Application
Publisher : PATTIMURA UNIVERSITY

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30598/barekengvol19iss4pp2467-2480

Abstract

This study investigates the application of a truncated spline nonparametric regression model for biresponse analysis of longitudinal data, focusing on modeling monthly stock prices specifically opening and closing prices of three private banks in Indonesia: Bank Mayapada, Bank Mega, and Bank Sinar Mas. The data used in this research are secondary data sourced from the website Id.Investing.com and monthly financial statement publications of three private banks in Indonesia. Longitudinal data, combining cross-sectional and time-series dimensions, are utilized to capture trends and patterns not detectable in traditional cross-sectional analysis. The truncated spline method is selected for its adaptability to nonlinear relationships and abrupt data behavior changes. The model incorporates three predictor variables traded stock volume, total assets, and total liabilities and evaluates their influence on stock prices. Assumptions of longitudinal data are validated using the Ljung-Box autocorrelation test, Bartlett’s sphericity test, and Pearson correlation. Results confirm significant within-subject correlations, independence between subjects, and strong interdependence between response variables. The optimal configuration is determined using Generalized Cross Validation (GCV), with up to three knots considered for segmentation. Weighted Least Squares (WLS) is employed for parameter estimation, accounting for within-subject correlations. Model evaluation based on Mean Absolute Percentage Error (MAPE) indicates high accuracy, with all MAPE values below 5%. The highest MAPE value is 4.41% for the closing price of Bank Mayapada, while the lowest is 2.65% for the opening price of the same bank. The segmentation analysis reveals that traded stock volume and total assets positively influence stock prices, while total liabilities exhibit a predominantly negative impact. The model is limited to internal financial indicators and does not include external macroeconomic factors such as interest rates or inflation. This study is the first to apply a biresponse truncated spline nonparametric regression approach to analyze stock prices of private banks in Indonesia by simultaneously modeling both opening and closing prices, providing a flexible and effective method for capturing complex patterns in longitudinal financial data.
IMPLEMENTASI PROGRAM PENGEMBANGAN MASYARAKAT PT TRISENSA MINERAL UTAMA MELALUI PELATIHAN BUDIDAYA DAN PENGGEMUKAN SAPI POTONG SERTA PEMBUATAN PAKAN SILASE DI DESA TANI BHAKTI KECAMATAN LOA JANAN 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 8 (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.v8i8.%p

Abstract

PT Trisensa Mineral Utama adalah perusahaan minerba yang beroperasi di sekitar Desa Batuah dan Tani Harapan, Kecamatan Loa Janan, Kabupaten Kutai kartanegara, Kalimantan Timur yang mempunyai tanggung jawab sosial dan lingkungan sesuai Undang-Undang Nomor 74 Tahun 2007 tentang Perseroan Terbatas yang mengaharuskan perusahaan perseroan terbatas untuk membina masyarakat sekitar agar dapat meningkatkan taraf kehidupannya dalam bentuk Corporate Social Responsbility (CSR). Oleh karena itu, Pengembangan Masyarakat adalah salah satu pendekatan yang dipilih dalam implementasi Corporal Social Responsibility (CSR). Kegiatan implementasi perusahaan dalam program Pengembangan Masyarakat yaitu pelatihan budidaya dan penggemukan sapi potong serta pembuatan pakan silase yang dilaksanakan pada tanggal 14 November 2024 di Desa Tani Bhakti dan di hadiri oleh kelompok tani Desa Batuah. Kegiatan ini bertujuan untuk memberikan pengetahuan mengenai budidaya dan penggemukan sapi serta memberikan pengetahuan mengenai bagaimana cara membuat pakan silase yang baik. Metode yang dilakukan dalam pelatihan ini yaitu dengan metode ceramah dan praktek langsung. Hasil dari kegiatan pelatihan ini yaitu antusiasnya para kelompok tani dalam mengikuti kegiatan pelatihan. Hal tersebut di lihat dari banyaknya peserta yang memberikan pertanyaan, dan pengalaman selama kegiatan penyampaian materi dan praktek langsung di laksanakan.
Perbandingan Arima dan Fuzzy Time Series Markov Chain Untuk Meramalkan Prediksi Hasil Panen Kopi (Studi Kasus Kabupaten Bengkulu Tengah Tahun 2012-2022) Saragih, Emiya Surabina Br; Gumay, Fridz Meryatdas; Fajriyanti, Meisya; Siregar, Samuel Eurico; Agwil, Winalia
Diophantine Journal of Mathematics and Its Applications Vol. 3 No. 2 (2024)
Publisher : UNIB Press

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33369/diophantine.v3i2.32049

Abstract

Penelitian ini bertujuan untuk membandingkan metode ARIMA dan Fuzzy Time Series Markov Chain dalam meramalkan prediksi hasil panen kopi pada studi kasus Kabupaten Bengkulu Tengah tahun 2012-2022. Langkah-langkah yang dilakukan dalam penelitian ini meliputi pengumpulan data, analisis data, pemodelan data, uji model, evaluasi model, dan kesimpulan. Model ARIMA digunakan untuk memodelkan komponen linear dari data hasil panen kopi, sedangkan metode Fuzzy Time Series Markov Chain digunakan untuk memodelkan komponen non-linear. Uji model dilakukan dengan membandingkan hasil peramalan dari kedua model dengan data aktual. Evaluasi dilakukan dengan menggunakan metrik evaluasi seperti Mean Absolute Percentage Error (MAPE) dan Root Mean Square Error (RMSE). Berdasarkan hasil penelitian, diperoleh kesimpulan bahwa model ARIMA memberikan hasil peramalan yang lebih baik dibandingkan dengan metode Fuzzy Time Series Markov Chain. Namun, perlu diingat bahwa pemilihan model terbaik tidak hanya ditentukan oleh nilai error dan akurasi saja. Terdapat beberapa faktor lain yang perlu dipertimbangkan, seperti karakteristik data, kompleksitas model, dan interpretasi hasil peramalan.
Perbandingan Metode Regresi Ridge dan Jackknife Ridge Regression pada Data Tingkat Pengangguran Terbuka Andini, Agita; Sunandi, Etis; Novianti, Pepi; Sriliana, Idhia; Agwil, Winalia
Limits: Journal of Mathematics and Its Applications Vol. 22 No. 1 (2025): Limits: Journal of Mathematics and Its Applications Volume 22 Nomor 1 Edisi Ma
Publisher : Pusat Publikasi Ilmiah LPPM Institut Teknologi Sepuluh Nopember

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12962/limits.v22i1.3374

Abstract

Regression analysis is a statistical technique used to analyze the relationship between predictor and response variables. One of the parameter estimation methods commonly used for regression analysis is Ordinary Least Squares. This method produces unbiased and efficient estimates, known as BLUE (Best Linear Unbiased Estimator). In multiple linear regression analysis involving more than one predictor variable, it is essential to meet model assumptions such as the absence of multicollinearity. Multicollinearity is a condition where predictor variables have a high correlation, which can disrupt the stability of parameter estimates. Therefore, Ridge Regression and Jackknife Ridge Regression methods were used to address this issue. Both methods modify the least squares method by adding a bias constant value. This research uses the Open Unemployment Rate (OUR) data in Sumatra in 2022, and 3 predictor variables exhibit multicollinearity. Based on the analysis comparing the Mean Squared Error (MSE) values, the Jackknife Ridge Regression method yields the smallest MSE value, 0.004. Both methods are effective in addressing multicollinearity and identifying significant predictor variables for OUR in Sumatra Island, namely the Human Development Index (HDI), average years of schooling, number of poor people, Life Expectancy (LE), population density and inactive population
Desa Cantik, Desa Cakap Statistik Agwil, Winalia; Sriliana, Idhia; Rini, Dyah Setyo; Supianti, Filo; Oktarina, Cinta Rizki; Famuji, Ahmad
Journal Of Human And Education (JAHE) Vol. 4 No. 1 (2024): Journal Of Human And Education (JAHE)
Publisher : Universitas Pahlawan Tuanku Tambusai

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31004/jh.v4i1.708

Abstract

Dalam pembangunan desa tentunya diperlukan pengetahuan terkait dengan potensi desa yang dimiliki. Pengembangan potensi desa dapat dilakukan dengan penggalian data awal serta pengumpulan data untuk pemetaan potensi desa. Sayangnya, sering kali terdapat desa yang memiliki sumber daya dengan kompetensi yang masih belum memadai sehingga di perlukannya peningkatan kompetensi perangkat desa tentang pengumpulan dan pemanfaatan data. Program studi S1 Statistika dan Pojok Statistik Universitas Bengkulu bermitra bersama BPS melakukan pengembangan peningkatan kompetensi terhadap perangkat desa dengan melakukan pengabdian Desa Cantik yang diharapkan dapat meningkatkan kemampuan perangkat desa dalam memanfaatkan data. Pengabdian ini dilakukan pada Kelurahan Tanah Panah Kota Bengkulu. Melalui pengabdian ini, diharapkan pihak terkait dapat memanfaatkan data yang ada. Hasil dari pengabdian ini mampu memberikan pemahaman terkait dengan pemanfaatan data desa serta pengolahan data Microsoft excel dan Canva. Dimana perangkat desa mengetahui tipe-tipe data dan pengaplikasiannya dalam memvisualisasi data pada Microsoft Excel, mengetahui fitur-fitur pada canva yang dapat digunakan pada elemen-elemen infografis. Sebagai rekomendasi bentuk kegiatan pengabdian Desa Cantik patut dilakukan untuk desa-desa yang lain untuk meningkatkan kemampuan dalam memanfaatkan data-data desa.
Negative Binomial Panel Regression Modeling on Amount of Crimes In Lampung Province Suratmin, Idam Abdurrohim Hasani; Agustina, Dian; Agwil, Winalia
EKSAKTA: Journal of Sciences and Data Analysis VOLUME 5, ISSUE 1, April 2024
Publisher : Fakultas Matematika dan Ilmu Pengetahuan Alam

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.20885/EKSAKTA.vol5.iss1.art4

Abstract

Crime in Lampung province is among the 10 highest in Indonesia in 2021. This study aims to obtain a model of the number of crimes and factors influencing it using negative binomial panel regression. The data used is in the form of panel data from the Lampung Province BPS website and publications for 2017-2021. The condition of data on the number of crimes as discrete and overdispersed data makes the negative binomial panel regression method more suitable than Poisson panel regression. Overdispersion is a state where the variance of the data is greater than the mean value of the data. Overdispersion causes the standard error (SE) of the estimated value to decrease, so that variables that should not be significant become significant. The factors thought to be the cause of crime are percentage of poverty (X1), population density (X2), expenditure per capita (X3), unemployment rate (X4), regional gross domestic income (X5), and the average duration of schooling (X6). The results of the analysis obtained for the selected panel data model are the negative binomial random effects (REBN), the influencing factors being X1, X3, X4 and X5. The districts/cities with the largest individual random effects were in the Way Kanan district and the smallest were in Metro City.
Meningkatkan Kinerja Model Klasifikasi Curah Hujan Melalui Penanggulangan Missing Value Dengan Imputasi Berbasis Model Agwil, Winalia; Agustina, Dian; Fransiska, Herlin; Hasani, Idam Abdurrohim
Innovative: Journal Of Social Science Research Vol. 4 No. 1 (2024): Innovative: Journal Of Social Science Research
Publisher : Universitas Pahlawan Tuanku Tambusai

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31004/innovative.v4i1.9158

Abstract

Penanggulangan terhadap data yang tidak lengkap telah banyak dikembangkan. Salah satu metode yang digunakan untuk menangani data hilang (missing value) adalah imputasi. Beberapa metode imputasi berbasis model untuk menduga nilai hilang pada variabel numerik adalah menggunakan metode regresi dan Random Forest. Penerapan metode imputasi ini dilakukan pada data curah hujan di Provinsi Bengkulu dari tahun 2013 sampai 2022 dengan variabel yang digunakan Y1 = Curah Hujan, X1 = Suhu Rata-rata, X2 = Suhu Maksimum, X3 = Suhu Minimum, X4 = Tekanan Udara diatas Permukaan Laut, X5 = Arah Angin, X6 = Kecepatan Angin Maksimum, X7 = Tingkat Awan, X8 = Lama Penyinaran Matahari. Hasil diperoleh bahwa kedua metode imputasi berbasis model yang digunakan memberikan nilai akurasi yang tidak terlalu berbeda. Nilai akurasi konsisten antara data training dan data testing, hal ini mengindikasikan bahwa imputasi yang dilakukan sudah baik. Pemodelan dilakukan dengan menggunakan metode CART dengan variabel dengan kontribusi tinggi adalah variabel yang memiliki kontribusi paling besar adalah arah angin pada hari sebelumnya dan variabel tutupan awan pada satu hari sebelumnya.