VARIANSI: Journal of Statistics and Its application on Teaching and Research

Transformasi Wavelet Diskrit Daubechies Fungsi Soft Thresholding untuk Prediksi Data Inflasi di Indonesia

2024-09-27T10:38:18+08:00

Metode prediksi klasik pada umumnya mensyaratkan data harus bersifat stasioner, akan tetapi umumnya terdapat beberapa data runtun waktu yang bersifat nonstasioner. Salah satu metode yang mampu menganalisa data non-stasioner dengan baik adalah transformasi wavelet diskrit (TWD). Estimasi thresholding dilakukan untuk menghapuskan noise pada data, dengan menggunakan fungsi dan parameter threshold yang sangat berpengaruh pada kemulusan hasil estimasi. Penelitian ini bertujuna untuk memprediksi inflasi di Indonesia pada bulan Mei 2012 hingga Desember 2022 menggunakan TWD daubechies soft thresholding dan parameter adaptive, serta mengetahui level resolusi terbaik dari perolehan nilai akurasi prediksi menggunakan Mean Absolute Percentage Error (MAPE). Hasil penelitian menunjukkan bahwa nilai prediksi baik pada level 1 sampai dengan level 5 sangat mendekati pola data aktual, dengan MAPE kurang dari 5%. Nilai MAPE terkecil yaitu 2,81% terdapat pada level resolusi pertama, dengan kategori akurasi prediksi sangat baik.

Rancangan Acak Kelompok (RAK) pada Analisis Pengaruh Pupuk Bokashi Kotoran Ayam terhadap Pertumbuhan dan Produksi Tanaman Bawang Daun

2024-09-27T10:38:18+08:00

This research focused on assessing the impact of chicken manure bokashi on the growth and yield of leeks and determining the optimal dosage for maximizing these outcomes. The study took place from June to August 2023 in Lasalepa Village, Lasalepa District, Muna Regency, using a Randomized Block Design (RBD) with a single factor: chicken manure bokashi fertilizer, tested at five different levels. These levels included a control with no bokashi (B0) and four varying doses: 5 tons/ha (0.6 kg/plot, B1), 10 tons/ha (1.2 kg/plot, B2), 15 tons/ha (1.8 kg/plot, B3), and 20 tons/ha (2.4 kg/plot, B4). Each treatment was repeated three times, resulting in a total of 15 experimental units. The parameters measured included plant height, the number of leaves per clump, the number of tillers per clump, and wet weight per clump. Data were analyzed using analysis of variance (ANOVA), and any significant effects were further explored using the Least Significant Difference (LSD) test at a 95% confidence level. The findings indicated that chicken manure bokashi had a highly significant effect on all measured parameters. The highest yield was observed with the application of 20 tons/ha (2.4 kg/plot) of bokashi, which resulted in an average wet weight of 27.49 grams per clump, equivalent to 3.66 tons/ha.

Implementasi Metode Support Vector Machine untuk Analisis Sentimen pada Ulasan Aplikasi Sayurbox di Google Play Store

2024-09-27T10:38:18+08:00

Penelitian ini bertujuan untuk menganalisis kinerja metode Support Vector Machine (SVM) dalam mengklasifikasikan ulasan pengguna aplikasi Sayurbox yang terkenal di Indonesia. Data ulasan diperoleh melalui scraping dari Google Play Store antara tahun 2017 hingga 2023. Ulasan dan rating yang diberikan pengguna digunakan sebagai indikator untuk mengevaluasi kepuasan terhadap layanan yang disediakan. Dalam penelitian ini, metode SVM digunakan untuk memproses data ulasan tersebut. Hasil klasifikasi menunjukkan bahwa metode SVM mencapai akurasi sebesar 89,29%. Selain itu, berdasarkan Confusion Matrix, nilai precision yang diperoleh adalah 91,42%, recall 95,58%, dan f1-score 93,50%. Temuan ini menunjukkan bahwa SVM merupakan metode yang efektif dalam mengklasifikasikan ulasan pengguna, yang dapat memberikan wawasan berharga untuk meningkatkan kualitas pelayanan Sayurbox.

Pengendalian Kualitas Statistika Pada Pupuk ZA III di PT Petrokimia Gresik Menggunakan Maximum Half Normal Multivariate Control Chart (Max-Half-Mchart)

2024-10-04T16:53:11+08:00

Max-Half-Mchart is one of the multivariate control diagrams that can monitor the average and variability of processes simultaneously or simultaneously. This diagram is a development of the Max-Mchat diagram with a half-normal distribution approach. This diagram is able to detect out of control signals well when there is a process shift. The data used in this study is data on the characteristics of ZA III product quality at PT Petrokimia Gresik. The quality control of ZA III Fertilizer at PT Petrokimia Gresik using Max-Half-Mchart shows that there is a shift in variability and average caused by the consistency of ( ) which is different in each production process, resulting in less than optimal products. Meanwhile, the overall production in the time period from October 2022 to September 2023 has been capable.

PEMODELAN GENERALIZED POISSON REGRESSION PADA KASUS AIDS DI PROVINSI NUSA TENGGARA TIMUR TAHUN 2023

2024-11-18T11:58:10+08:00

East Nusa Tenggara (ENT) Province is one of the contributors to the high number of AIDS cases in Indonesia. The purpose of study is to analyse the factors affecting the number of AIDS cases in the Province using the Generalized Poisson Regression method. The variable used is the number AIDS cases as a response variable, and seven predictor variables that might be influence the response variable. The data obtained from the publications of the Central Statistics Agency of ENT Province in 2024. The study shows that the GPR model applying six predictors yields the best model with value of AIC is 419.93. Those variables are the number of health facilities , the number of poor people , the human development index , sexually transmitted infections , the number of adolescents aged 15-24 years who received reproductive health counselling , and the open unemployment rate