Claim Missing Document
Check
Articles

Found 4 Documents
Search
Journal : Inferensi

Penerapan Keluarga Model Spline Truncated Polinomial pada Regresi Nonparametrik Andrea Tri Rian Dani; Ludia Ni’matuzzahroh
Inferensi Vol 5, No 1 (2022): Inferensi
Publisher : Department of Statistics ITS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12962/j27213862.v5i1.12537

Abstract

One approach that is often used by researchers to determine the form of the relationship pattern between the response variables and predictor variables in regression analysis, namely the nonparametric approach, where the approach is used when the shape of the regression curve is assumed to be unknown. The truncated spline is a polynomial model in nonparametric regression that has segmented properties, where these properties provide better flexibility than ordinary polynomial models and are able to handle data whose behavior changes in certain sub-intervals due to the knot points in it. This study aims to apply a family of spline truncated polynomial models to nonparametric regression in the case of automotive data. The estimation method used is Ordinary Least Square (OLS). The number of knot points tested is 1 to 4-knot points with a degree of p=1,2,3. Based on the results of the analysis, the best model that produces the smallest GCV value is the nonparametric spline truncated quadratic regression model with 4 knots, which produces a GCV value of 522.27 and a coefficient of determination of 79.77%.
Pemodelan Regresi Nonparametrik Spline Truncated pada Data Longitudinal Andrea Tri Rian Dani; Ludia Ni’matuzzahroh; Vita Ratnasari; I Nyoman Budiantara
Inferensi Vol 4, No 1 (2021): Inferensi
Publisher : Department of Statistics ITS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12962/j27213862.v4i1.8737

Abstract

Saat ini pendekatan regresi nonparametrik banyak mendapat perhatian dari para peneliti, dikarenakan memiliki fleksibilitas yang tinggi dan tidak tergantung pada asumsi bentuk kurva regresi. Spline truncated adalah salah satu model dalam regresi nonparametrik yang sering digunakan, karena mampu menangani data yang perilakunya berubah-ubah pada sub-sub interval tertentu. Pada analisis regresi, data yang seringkali digunakan adalah data cross-section, namun yang sebenarnya adalah analisis regresi juga dapat diterapkan pada data longitudinal, khususnya dengan menggunakan pendekatan regresi nonparametrik. Data longitudinal merupakan gabungan antara data cross-section dan time series. Penelitian ini bertujuan untuk memodelkan data persentase penduduk miskin di Provinsi Papua Tahun 2016 hingga Tahun 2019 menggunakan model regresi nonparametrik spline truncated. Metode estimasi parameter yang digunakan adalah Weighted Least Squares (WLS). Banyaknya titik knot yang dicobakan adalah 1 hingga 3 titik knot. Berdasarkan hasil analisis, model regresi nonparametrik spline truncated terbaik adalah model yang menggunakan 1 titik knot, dengan nilai GCV yang paling minimum yaitu sebesar 8,05 dan Koefisien Determinasi (R2) sebesar 99,98%.
Aplikasi Model ARIMAX dengan Efek Variasi Kalender untuk Peramalan Trend Pencarian Kata Kunci “Zalora” pada Data Google Trends Andrea Tri Rian Dani; Sri Wahyuningsih; Fachrian Bimantoro Putra; Meirinda Fauziyah; Sri Wigantono; Hardina Sandariria; Qonita Qurrota A'yun; Muhammad Aldani Zen
Inferensi Vol 6, No 2 (2023)
Publisher : Department of Statistics ITS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12962/j27213862.v6i2.15793

Abstract

ARIMAX is a method in time series analysis that is used to model an event by adding exogenous variables as additional information. Currently, the ARIMAX model can be applied to time series data that has calendar variation effects. In short, calendar variations occur due to changes in the composition of the calendar. The purpose of this study is to apply the ARIMAX model with the effects of calendar variations to forecast search trends for the keyword "Zalora". Data were collected starting from January 2018 to November 2022 in the form of a weekly series. Based on the results of the analysis, the ARIMAX model is obtained with calendar variation effects with ARIMA residuals (1,1,1). Forecasting accuracy using the Mean Absolute Percentage Error (MAPE) of 10.47%. Forecasting results for the next 24 periods tend to fluctuate and it is estimated that in April 2023 there will be an increase in search trends for the keyword "Zalora".
Aplikasi Pengelompokan Data Runtun Waktu dengan Algoritma K-Medoids Muhammad Aldani Zen; Sri Wahyuningsih; Andrea Tri Rian Dani
Inferensi Vol 6, No 2 (2023)
Publisher : Department of Statistics ITS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12962/j27213862.v6i2.15864

Abstract

The development of information technology will always be accompanied by the storage and accumulation of massive quantities of digital information. Cluster analysis is one of many data processing problems that require the selection of an appropriate algorithm when dealing with large data sets. Cluster analysis is a collection of techniques for dividing a set of observation objects into clusters. Cluster analysis is applicable to time series data, the processing of which differs slightly from that of cross-section data. Clustering time series is a technique for processing multivariable time series data. K-Medoids is the clustering algorithm used for time series clustering. The objective of this study is to obtain optimal K-values in determining the number of clusters based on silhouette coefficients and grouping outcomes using the K-Medoids algorithm. In this study, the dynamic time-warping distance is utilized as the similarity metric. This study provides cooking oil price data for 34 Indonesian provinces from October 2017 to October 2022. The optimal K value is determined for two clusters based on the results of the analysis, with 19 provinces joining cluster 1, where the cluster with cooking oil prices was below cluster 2 and 15 provinces joining cluster 2 which is the cluster with the highest cooking oil prices.
Co-Authors A'yun, Qonita Qurrota Adhitya Ronnie Effendie, Adhitya Ronnie AINURROCHMAH, ALIFTA Alifta Ainurrochmah Alifta Ainurrochmah Anisar, Anggi Putri AVIANTHOLIB, IGAR CALVERIA Avrilia, Khairunnisa Budi Cahyono Budi, Ennesya Estya Candra, Yossy Chandra, Yossy Dandito Laa Ull Darnah Darnah, Darnah Dimas Nugroho Dwi Seputro Fachrian Bimantoro Putra Fadlirhohim, Rizki Dwi Fauziyah, Meirinda Fidia Deny Tisna Amijaya Goenjatoro, Rito Gunardi Gunardi Hardina Sandariria Hinadang, Elen A. I Gusti Bagus Ngurah Diksa I Nyoman Budiantara I Nyoman Budiantara Ibaad, Muhammad Irsadul indarsih, Indarsih Koirudin, Hadi Kosasih, Raditya Arya Krisna Rendi Awalludin Ludia Ni'matuzzahroh Ludia Ni’matuzzahroh M. Fathurahman Mahmuda, Siti Mar’ah, Zakiyah Meirinda Fauziyah Melisa Arumsari Memi Nor Hayati Mislan Muawanah, Chusnul Muhammad Aldani Zen Mulyadi, Taqriri Kamal Nanda Arista Rizki NARITA YURI ADRIANINGSIH Ni'matuzzahroh, Ludia Nilam Novita Sari Novidianto, Raditya Nurul Istiqomah Oroh, Chiko Zet Puspitasari, Melda Putra, Fachrian Bimantoro Qonita Qurrota A'yun Raditya Arya Kosasih Raditya Novidianto Rahayu, Joana K. Rahmah, Syifa M. Rahmah, Syifa Mutia Rahmania Rahmania Ramadhani, Bagus D. Regita Putri Permata Rifdatun Ni’mah Riry Sriningsih Rito Goejantoro, Rito Sifriyani, Sifriyani Siringoringo, Meiliyani Sitinjak, Jesselin Paskalis Solikhah, Arifatus Solikhatun, Solikhatun Sri Wahyuni Sri Wahyuningsih Sri Wahyuningsih Sri Wigantono Surya Prangga Suyitno Suyitno Syaripuddin Syaripuddin Tanur, Erwin Tutik Handayani, Tutik Uha Isnaini Vita Ratnasari Wahyujati, Mohamad Fahruli Watika, Noor Hikmah Zen, Muhammad Aldani