Miftahul Irfan

Abstract High-dimensional panel vector autoregressive (PVAR) models have become increasingly important for analyzing complex dynamic interactions in multivariate time series systems. However, conventional estimation methods often suffer from overparameterization, multicollinearity, and poor forecasting performance, particularly when the number of parameters grows rapidly relative to the sample size. To address these challenges, this study proposes a Lag-Adaptive Hierarchical Sparse Bayesian Panel Vector Autoregressive (LAHS-BPVAR) model that integrates adaptive lag regularization, hierarchical global-local shrinkage, and sparse Bayesian learning within a unified framework. The proposed model introduces a novel prior covariance structure in which coefficient variances decay according to a hybrid lag-adaptive mechanism combining polynomial and exponential shrinkage, allowing automatic penalization of distant lags while preserving relevant short-run dynamics. A hierarchical sparse prior based on global-local shrinkage further enables adaptive variable selection in high-dimensional settings. Theoretical analysis establishes posterior contraction and sparse support recovery consistency under standard regularity conditions, providing formal asymptotic justification for the proposed estimator. Estimation is performed using a Gibbs sampler with Metropolis-Hastings updates for non-conjugate hyperparameters. Monte Carlo simulation results demonstrate competitive estimation and forecasting performance relative to conventional VAR, Minnesota Bayesian VAR, LASSO VAR, and standard Horseshoe Bayesian VAR. The proposed framework consistently achieves improved sparse support recovery through more effective control of false positive coefficient selection while maintaining estimation accuracy. An empirical application to Indonesian macroeconomic time series further demonstrates the practical effectiveness of the proposed framework in capturing dynamic economic relationships while maintaining model parsimony. The findings suggest that the proposed methodology provides a robust and flexible solution for high-dimensional panel time series modeling.

Kemampuan dalam memanfaatkan teknologi, khususnya dalam eksplorasi dan visualisasi data, menjadi tantangan signifikan dalam dunia pendidikan di era digital saat ini. SMKS Nurul Huda Pringsewu sebagai mitra kegiatan pengabdian menunjukkan adanya keterbatasan kompetensi siswa dalam mengolah data dan menyajikannya dalam bentuk visual yang informatif. Menanggapi permasalahan tersebut, tim Pengabdian kepada Masyarakat (PkM) melaksanakan pelatihan menggunakan aplikasi Looker Studio untuk eksplorasi data dan pembuatan dashboard statistik. Looker Studio merupakan platform berbasis web yang memudahkan pengguna dalam mengolah data numerik dan menyajikannya secara visual, sehingga proses pembacaan data tidak lagi dilakukan secara manual. Pelatihan ini bertujuan untuk meningkatkan kemampuan siswa dalam mengolah dan memvisualisasikan data menggunakan teknologi terkini. Hasil dari kegiatan ini menunjukkan peningkatan kompetensi siswa dalam mengeksplorasi data statistik serta peningkatan kualitas pembuatan dashboard yang lebih informatif dan menarik. Kegiatan ini diharapkan dapat menjadi langkah awal dalam membekali siswa dengan keterampilan digital yang relevan dengan kebutuhan industri saat ini.

Students' misconceptions in estimating the parameters of probability distribution are still an important problem in learning statistics in universities. Most previous studies have examined partial misconceptions from cognitive aspects, so the structural relationship between concept understanding, learning experience, learning motivation, and problem-solving skills in explaining misconceptions has not been widely analyzed in an integrated manner. This study aims to develop and validate a structural model that explains the factors that influence student misconceptions in estimating probability distribution parameters using the Partial Least Squares Structural Equation Modeling (PLS-SEM) approach. This study uses an explanatory quantitative design involving 200 students who have studied probability distribution and parameter estimation in several universities in Lampung Province. Data was collected through a Likert scale questionnaire that measured five latent constructs, namely concept comprehension, learning experience, learning motivation, problem-solving skills, and student misconceptions. The analysis was carried out through the evaluation of the measurement model (loading factor, composite reliability, and average variance extracted), structural model testing using the bootstrapping technique, and evaluation of the overall suitability of the model. The results showed that concept comprehension (β = 0.74) and learning experience (β = 0.82) had a significant effect on problem-solving skills. Problem-solving skills further affect learning motivation (β = 1.92), while learning motivation affects the level of student misconception (β = 0.67). The developed model was able to explain 65% of the variation in problem-solving skills and 88% of the variation in student misconceptions. This research contributes in the form of a SEM-PLS model that integrates cognitive and affective factors in explaining the emergence of statistical misconceptions, as well as providing an empirical basis for the development of more effective statistical learning strategies to reduce student misconceptions.

This research aims to calculate agricultural insurance premiums for rice crops in Lampung Province using the Black–Scholes model approach. The background of the study lies in the high production risks caused by climate variability, extreme weather, flooding, drought, and pest attacks, all of which significantly affect the potential for crop failure. Secondary data in the form of harvested area and rice production for the 2019–2023 period were obtained from official publications of the Statistics Indonesia. The analysis shows that rice yield per hectare follows a lognormal distribution based on the Kolmogorov–Smirnov test, thus meeting the fundamental assumptions of the model. Calculation parameters, including production volatility, risk-free interest rate, threshold value, and production time, were used in estimating the premium. The results indicate that the probability of production falling below the threshold is 13.69%, while the calculated premium is IDR 802,584 per planting season. These findings confirm that the Black–Scholes model can be applied as an objective quantitative approach to determining risk-based agricultural insurance premiums. The application of this method has the potential to improve the accuracy of premium calculations and support the development of a fairer and more sustainable Rice Farming Insurance Scheme (AUTP).

Health data are often analyzed in their continuous form through approaches such as linear, logistic, or survival models. In this study, hematological variables were dichotomized based on established clinical cut-offs to enable log-linear analysis of associations among categorical variables, acknowledging the potential loss of information from this transformation. A log-linear model was applied to evaluate independence, dependence, and interaction patterns among leukocyte, hemoglobin, and hematocrit categories in a dengue hemorrhagic fever (DHF) patient dataset. Previous analyses using survival models identified these variables as factors associated with recovery rates; however, these models did not capture their interaction structure. Log-linear analysis was therefore employed to examine these associations more comprehensively. The best-fitting model was identified as , which included two-factor interactions between leukocyte–hematocrit and hemoglobin–hematocrit. This model demonstrated a good fit (Pearson , , ), including a three-factor interaction resulted in a saturated model (= 0) and did not improve model performance. These findings highlight significant interaction patterns among hematological variables in DHF patients, providing a more detailed understanding of their joint associations.

Penelitian ini bertujuan untuk menganalisis pengaruh faktor-faktor ekonomi dan sosial terhadap tingkat kemiskinan di Provinsi Lampung pada tahun 2023. Variabel independen yang dikaji meliputi Upah Minimum Provinsi (UMP), Indeks Pembangunan Manusia (IPM), Pertumbuhan Ekonomi, dan Tingkat Pengangguran Terbuka (TPT). Metode analisis yang digunakan adalah regresi linear berganda untuk mengidentifikasi dan mengukur pengaruh simultan variabel-variabel tersebut terhadap tingkat kemiskinan sebagai variabel dependen. Data yang digunakan merupakan data sekunder yang diperoleh dari Badan Pusat Statistik (BPS) Provinsi Lampung dan instansi terkait. Hasil penelitian menunjukkan bahwa UMP, IPM, dan Pertumbuhan Ekonomi berpengaruh signifikan terhadap tingkat kemiskinan, sedangkan TPT tidak menunjukkan pengaruh yang signifikan. Temuan ini memberikan wawasan penting mengenai determinan kemiskinan di Provinsi Lampung dan dapat dijadikan dasar dalam perumusan kebijakan yang berorientasi pada peningkatan kesejahteraan masyarakat melalui perbaikan upah minimum, pengembangan kualitas sumber daya manusia, dan penguatan ekonomi daerah

This Community Service (PkM) program aims to enhance students’ competencies in data analysis and statistical dashboard management through the utilization of the Looker Studio application. The training was conducted at SMKS Nurul Huda Pringsewu, involving students as participants. The training methods included lectures, demonstrations, and hands-on practice in processing data and presenting it in the form of interactive dashboards. The results of the program showed that students were able to understand the basic concepts of data exploration, the purpose of data visualization, and the use of key features in Looker Studio. In addition, students’ skills in selecting appropriate chart types according to analytical needs improved significantly. Based on the satisfaction survey, most participants rated the activity as very satisfactory (63%) and satisfactory (16%), although a small proportion expressed dissatisfaction (16%) or were not satisfied (5%). Overall, this PkM activity successfully contributed to improving students’ data literacy and digital skills, which are expected to support them in facing both academic challenges and the demands of a data-driven workforce

In graph theory, the locating chromatic number refers to the minimum number of colours required to colour the vertices of a graph such that each vertex is uniquely identifiable by its own colour and the colours of its neighbouring vertices. This concept is associated with graph coloring, which entails assigning colors to a graph's vertices in such a way that no two adjacent vertices have the same color. It determines the minimum number of colours necessary to produce a proper vertex colouring. This research focuses on determining the locating chromatic number of Pizza graphs containing a single path on the outer vertices.

Permasalahan dalam penelitian ini adalah tingginya jumlah variabel yang saling berkorelasi, sehingga menyulitkan pemahaman terhadap struktur data. Tujuan penelitian ini untuk mereduksi dimensi variabel yang saling berkorelasi dan memperoleh pemahaman yang lebih baik terhadap struktur data. Data yang digunakan terdiri dari 768 sampel dengan 8 variabel bebas dan 1 variabel terikat pada Data Diabetes. Langkah-langkah analisis meliputi penentuan jumlah komponen utama, uji Bartlett dan uji Keiser-Meyer-Olkin (KMO) untuk memastikan kecocokan data, perhitungan koefisien komponen utama, serta visualisasi grafik AKU. Hasil analisis menunjukkan bahwa terdapat 5 komponen utama yang mampu menangkap lebih dari 80% keragaman data, serta hubungan yang beragam antar variabel yang diamati. The problem in this research is the high number of variables that weaken each other, making it difficult to understand the data structure. The aim of this research is to reduce the dimensions of mutually burdening variables and gain a better understanding of the data structure. The data used consists of 768 samples with 8 independent variables and 1 dependent variable in Diabetes Data. The analysis steps include determining the number of principal components, Bartlett's test and Keiser-Meyer-Olkin (KMO) test to ensure data suitability, performance of principal component coefficients, and visualization of the AKU graph. The results of the analysis show that there are 5 main components that are able to capture more than 80% of the diversity of the data, as well as various relationships between the observed variables.

To analyze economic data, it would be better to use panel data (a combination of time series and cross section) because there is a lot of information obtained. One model that is widely used in panel data analysis is the PVAR model because it was able to identify the dynamic effects of the heterogeneity shock of the unobserved variables and their causal effects. PVAR is widely used in data on the economic growth of a region or country. Many factors affect economic growth, one of which is inflation. Inflation and economic growth are two things that influence each other. Economic growth it self can be calculated based on the value of GDP. In this study, Granger causality shows that there is a two-way influence of the two variables. The PVAR model shows that the two variables influence each other but the time effect has a greater influence. Where the GDP rate in the previous year had more influence on the current GDP rate than the inflation rate.

Diabetes merupakan salah satu penyakit yang menjadi penyebab kematian terbesar didunia. Kasus kematiannya pun tercatat lebih dari 4 juta pada tahun 2019. Diabetes juga dapat menyebabkan timbulnya penyakit lainnya. Bahaya diabetes ini menjadi perhatian khusus WHO. Seiring dengan perkembangan teknologi ini, banyak sekali kolaborasi antara bidang kesehatan, statistic dan computer untuk menanggulangi berbagai macam penyakit. Algortima machine learning menjadi popular dalam proses klasifikasi data dan sudah banyak diterapkan pada data kesehatan. Dengan begitu pada artikel ini akan dilakukan perbandingan algoritma machine learning KNN, Decision Tree, dan SVM untuk melihat algortima mana yang paling cocok untuk klasifikasi data diabetes. Hasil menunjukkan bahwa KNN dan SVM memiliki akurasi yang cukup besar yaitu 81,13%. Sehingga kedua algortima tersebut dapat menjadi rekomendasi proses klasifikasi data diabetes sehingga dapat membantu dokter dalam menanggulangi penyakit diabetes. Hasil ini juga menunjukkan bahwa 8 variabel yang digunakan berpengaruh terhadap resiko diabetes