Articles
Data Non-normality on AMMI Models: Box-Cox Transformations
Alfian Futuhul Hadi;
Halimatus Sa'diyah;
I Made Sumertajaya
Jurnal ILMU DASAR Vol 8 No 2 (2007)
Publisher : Fakultas Matematika dan Ilmu Pengetahuan Alam Universitas Jember
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
Full PDF (490.603 KB)
AMMI (Additive Main Effect Multiplicative Interaction) model for interactions in two-way table provide the major mean for studying stability and adaptability through genotype × environment interaction (GEI), which modeled by full interaction model. Eligibility of AMMI models depends on that assumption of normally independent distributed error with a constant variance. In the case of non-normal data distribution, the appropriateness of AMMI model is being doubtful. Transform the observation by power family of Box-Cox transformation is an effort to handle the non-normality. AMMI model then can be applied to the transformed data appropriately following by the use of ordinary least square for estimating parameters. This approach is investigated by applying them to (i) a count data of pest population of Poisson distribution, which came from a study of leave pest in soybean genotype, and to (ii) a study of rice genotype stability of filled grain per panicle (Binomial data). One must be carefully considered what the meaning of the transformation in the AMMImodels and Biplot AMMI.
Generalized AMMI Models for Assessing The Endurance of Soybean to Leaf Pest
Alfian Futuhul Hadi;
A. A. Mattjik;
IM Sumertajaya
Jurnal ILMU DASAR Vol 11 No 2 (2010)
Publisher : Fakultas Matematika dan Ilmu Pengetahuan Alam Universitas Jember
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
Full PDF (217.038 KB)
AMMI(Additive Main Effect Multiplicative Interaction)model for interactions in two-way table provide the major mean for studying stability and adaptability through genotype × environment interaction (GEI), which modeled by full interaction model. Eligibility of AMMI (Additive Main Effect Multiplicative Interaction) model depends on that assumption of normally independent distributed error with a constant variance. In the study of genotypes’ resistance, disease and pest (insect) incidence on a plant for example, the appropriateness of AMMI model is being doubtful. We can handle it by introducing multiplicative terms for interaction in wider class of modeling, Generalized Linear Models. Its called Generalized AMMI model. An algorithm of iterative alternating generalized regression of row and column estimates its parameters. GAMMI log-link model will be applied to the Poisson data distribution. GAMMI log-link models give us good information of the interaction by its log-odd ratio.
Sentiment Analysis on Covid-19 Vaccination in Indonesia Using Support Vector Machine and Random Forest
I Made Sumertajaya;
Yenni Angraini;
Jamaluddin Rabbani Harahap;
Anwar Fitrianto
JUITA : Jurnal Informatika JUITA Vol. 10 No. 1, May 2022
Publisher : Department of Informatics Engineering, Universitas Muhammadiyah Purwokerto
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
Full PDF (1277.257 KB)
|
DOI: 10.30595/juita.v10i1.12394
World Health Organization (WHO) stated Covid-19 as a global pandemic in March, 2020. This pandemic has influenced people’s life in many sectors such as the economy, health, tourism, and many more. One way to end this pandemic is to make herd immunity obtained through the vaccination program. This program still raises pros and cons at the beginning of its implementation in Indonesia. Many people doubt the safety and side effects of the vaccine. There are also pros and cons to vaccination programs in social media such as Twitter. This platform generates a huge amount of text data containing people's perceptions about vaccines. This research aims to predict sentiment using supervised learning such as support vector machine (SVM) and random forest and capture sentiment about vaccines in Indonesia in the first two weeks of the program. The result shows SVM was a better model than random forest based on the precision and F1-score metrics. The SVM approach produces a precision value of 0.50, a recall of 0.64, and an F1-score of 0.52. In the study, it was also found that tweets with neutral sentiment dominated the twitter user sentiment in the study period. Tweets with negative sentiment decreased after the first week of the COVID-19 vaccination program.
Pemilahan Volatilitas Harga Daging Sapi Menggunakan Metode Ensemble Empirical Mode Decomposition
Fitria Hasanah;
Hari Wijayanto;
I Made Sumertajaya
Jurnal Agro Ekonomi Vol 38, No 1 (2020): Jurnal Agro Ekonomi
Publisher : Pusat Sosial Ekonomi dan Kebijakan Pertanian
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21082/jae.v38n1.2020.41-54
EnglishStaple food prices include the major determinants of households food security and general inflation. Beef is a basic food which its price is controlled by the Government of Indonesia. This study aims to identify the determinants beef price volatility using the Ensemble Empirical Mode Decomposition (EEMD) method. The data was a weekly series of Januari 2006–Desember 2018 obtained from the Ministry of Trade. EEMD extracts data into a number of Intrinsic Mode Functions (IMFs) that are independent which are then used to forecast beef prices with the ARIMA model. EEMD produced 6 IMFs and one residual. The residual contributed 99.85% to beef price volatility. This means that the long-term trend of beef prices is determined by the residual trends. The EEMD results indicate that the high beef price volatility in certain periods is mainly due to high demand during the Ramadhan month and Idul Fitri, import quota policy, and changes in exchange rates and petroleum prices. The IMF and residual based ARIMA forecasting model obtained MAPE value of 0.42% but with contradicting directions. The Government may use the import quota as a policy instrument for stabilizing the beef price.IndonesianHarga pangan pokok termasuk faktor penentu utama ketahanan pangan rumah tangga dan inflasi umum. Daging sapi adalah salah satu bahan pangan pokok yang harganya dikendalikan Pemerintah Indonesia. Penelitian ini bertujuan mengidentifikasi faktor penentu volatilitas harga daging sapi dengan metode Ensemble Empirical Mode Decomposition (EEMD). EEMD menguraikan data menjadi sejumlah Intrinsic Mode Function (IMF) yang saling bebas yang selanjutnya digunakan untuk melakukan peramalan harga daging sapi dengan model ARIMA. Data yang digunakan adalah harga daging sapi mingguan Januari 2006–Desember 2018 yang diperoleh dari Kementerian Perdagangan. EEMD menghasilkan 6 IMF dan satu sisaan. Sisaan IMF memberikan kontribusi sebesar 99,85% terhadap pergerakan harga daging sapi. Artinya bahwa tren jangka panjang harga daging sapi ditentukan oleh tren sisaan. Berdasarkan hasil EEMD, volatilitas harga daging sapi yang tinggi pada periode-periode tertentu dipengaruhi oleh beberapa faktor terutama tingginya permintaan selama bulan Ramadhan dan Idul Fitri dan kebijakan kuota impor, serta perubahan nilai tukar rupiah dan harga BBM. Model peramalan ARIMA yang diduga berdasarkan IMF dan sisaan IMF menghasilkan nilai MAPE sebesar 0,42%, namun arah perubahannya tidak bersesuaian. Disarankan agar pemerintah menggunakan kuota impor sebagai salah satu instrumen kebijakan stabilisasi harga daging sapi.
Geo-additive Models in Small Area Estimation of Poverty
Novi Hidayat Pusponegoro;
Anik Djuraidah;
Anwar Fitrianto;
I Made Sumertajaya
Journal of Data Science and Its Applications Vol 2 No 1 (2019): Journal of Data Science and Its Applications
Publisher : Telkom University
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/jdsa.2019.2.15
Spatial data contains of observation and region information, it can describe spatial patterns such as disease distribution, reproductive outcome and poverty. The main flaw in direct estimation especially in poverty research is the sample adequacy fulfilment otherwise it will produce large estimate parameter variant. The Small Area Estimation (SAE) developed to handle that flaw. Since, the small area estimation techniques require “borrow strength” across the neighbor areas thus SAE was developed by integrating spatial information into the model, named as Spatial SAE. SAE and spatial SAE model require the fulfilment of covariate linearity assumption as well as the normality of the response distribution that is sometimes violated, and the geo-additive model offers to handle that violation using the smoothing function. Therefore, the purpose of this paper is to compare the SAE, Spatial SAE and Geo-additive model in order to estimate at sub-district level mean of per capita income of each area using the poverty survey data in Bangka Belitung province at 2017 by Polytechnic of Statistics STIS. The findings of the paper are the Geo-additive is the best fit model based on AIC, and spatial information don't influence the estimation in SAE and spatial SAE model since they have the similar estimation performance.
KERAGAAN REGRESI LS, LAD, DAN MLAD PADA DATA DELIVERY TIME (The Performance of LS, LAD, and MLAD Regression on Delivery Time Data)
Setyono Setyono;
IM Sumertajaya;
A Kurnia;
AA Mattjik
JURNAL AGRONIDA Vol. 2 No. 1 (2016)
Publisher : Universitas Djuanda Bogor
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
Full PDF (984.471 KB)
|
DOI: 10.30997/jag.v2i1.750
Pendugaan koefisien regresi berbasis optimasi sisaan yang dikenal adalah dengan carameminimumkan jumlah kuadrat sisaan (LS) dan meminimumkan jumlah sisaan mutlak (LAD).Pendugaan dengan cara meminimumkan maksimum sisaan mutlak (MLAD) belum dikembangkan.Tujuan penelitian ini adalah untuk mengetahui apakah program linier dapat digunakan untukmendapatkan penduga koefisien regresi yang meminimumkan maksimum sisaan mutlak danmembandingkan hasilnya dengan hasil pendugaan menggunakan metode LS dan LAD. Data yangdigunakan adalah data Delivery Time yang biasa digunakan untuk uji coba metode regresi. Hasilpenelitian menunjukkan bahwa program linier dapat digunakan untuk mendapatkan pendugakoefisien regresi yang meminimumkan maksimum sisaan mutlak, pada data Delivery Time regresiLAD paling baik menurut kriteria validasi silang, sedangkan regresi LS paling stabil menurutsemua kriteria. Dalam metode MLAD dimungkinkan diperoleh subset pengamatan yangmenghasilkan penduga koefisien regresi yang sama besar dengan penduga koefisien regresi darikeseluruhan pengamatan.Kata kunci : MLAD, program linier, regresi, sisan mutlak, validasi silang
Research in Pesantren-based Higher Education: Exploring The Factors Improving Lecture's Research Performace
Abdu Alifah;
M. Syamsul Maarif;
I Made Sumertajaya
AL-ISHLAH: Jurnal Pendidikan Vol 14, No 2 (2022): AL-ISHLAH: Jurnal Pendidikan
Publisher : STAI Hubbulwathan Duri
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
Full PDF (580.751 KB)
|
DOI: 10.35445/alishlah.v14i2.1590
Integrating two different systems between higher education and pesantren (Islamic boarding schools), Pesantren-based Higher Education (PTBP) is alleged as an ideal Islamic higher education model considering its critical role in the process of integration-interconnection of science and Islamic teaching. Like other Islamic higher educations in Indonesia, PTBP represents poor organizational performance, particularly in research productivity and quality. Since research performance has the most considerable portion in the higher education performance evaluation both nationally and internationally, this study attempted to explore the factors that possibly improve lecturers' research performance at PTBP, categorized into the level of the individual, institutional and resource factors. This study involved 84 lecturers at 7 PTBP in Bogor, Indonesia, selected based on the stratified random sampling and voluntary sampling method. The collected data was analyzed using Structural Equation Modeling-Partial Least Square (SEM-PLS). The results showed that only research competence positively affected research performance at the individual level, but not for work motivation and time management. Research culture and academic leadership insignificantly affected research performance. However, only high-performance work practice (HPWP) at the institutional level positively affected research performance. Meanwhile, collaboration networks at the resource level positively affected research performance, but not for university research funding. The managerial implications are further discussed in this study.
PENGARUH ORGANISASI PEMBELAJAR DAN INOVASI TERHADAP PENINGKATAN KINERJA UKM DI KOTA BOGOR
Manuel Leonard Sirait;
Anggraini Sukmawati;
I Made Sumertajaya
Jurnal Manajemen Vol. 19 No. 2 (2015): June 2015
Publisher : Fakultas Ekonomi dan Bisnis, Universitas Tarumanagara
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.24912/jm.v19i2.127
Permasalahan yang dihadapi oleh UKM salah satunya ada keterbatasan sumberdaya manusia, yaitu kurangnya pengetahuan dan informasi yang dimiliki sehingga berpengaruh terhadap kinerja UKM. Menghadapi Masyarakat Ekonomi ASEAN akan mempengaruhi daya tahan UKM Kota Bogor karena persaingan yang semakin ketat dan melakukan perbaikan kualitas sumberdaya manusia pada UKM di Kota Bogor. Penelitian ini bertujuan untuk menganalisis pengaruh dari organisasi pembelajar dan inovasi terhadap kinerja UKM Kota Bogor. Penelitian ini dilakukan di Bogor dengan jumlah sampel UKM sebanyak 46 UKM dan sampel responden berjumlah 149 orang. Teknik pengambilan sampel dilakukan dengan menggunakan teknik multi stage sampling yaitu gabungan dari teknik stratified random sampling dan proportional sampling. Pengolahan data menggunakan analisis Structural Equation Modelling (SEM) dengan pendekatan Partial Least Square (PLS). Hasil analisis pada penelitian ini menunjukkan bahwa (1) terdapat pengaruh yang siginifkan antara organisasi pembelajar terhadap inovasi, (2) organisasi pembelajar memberikan pengaruh yang signifikan terhadap kinerja UKM, (3) inovasi memberikan pengaruh yang signifkan terhadap kinerja UKM Kota Bogor.The lack of human resources is one of the problem that small and medium enterprises (SMEs) has faced. Human resources in SMEs dont have sufficient knowledge and information and it effects the performances of SMEs. Dealing with people of ASEAN Economic Community will effect the resistance of SMEs in Bogor, due to the competitive and making an increase in quality of human resources towards SMEs in Bogor. This research was to analyze the effect of the learning organization and innovation towards performance of SMEs in Bogor. This research was conducted in Bogor with 46 samples for SMEs and 149 for respondent samples. The technique of sampling was conducted by using multi stage sampling namely the combination of stratified random sampling and proportional sampling. The tabulation data was analyzed by Structural Equation Modelling (SEM) by approaching of Partial Least Square (PLS). The result of this research were (1) the existence of significant influence between learning organization toward innovation, (2) the learning organization gives a significant effect towards the performance of SMEs, (3) innovation gives a significant effect towards the performance of SMEs in Bogor.
Pattern Detection of Economic and Pandemic Vulnerability Index in Indonesia Using Bi-Cluster Analysis
Wiwik Andriyani Lestari Ningsih;
I Made Sumertajaya;
Asep Saefuddin
JUITA : Jurnal Informatika JUITA Vol. 10 No. 2, November 2022
Publisher : Department of Informatics Engineering, Universitas Muhammadiyah Purwokerto
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
Full PDF (1361.823 KB)
|
DOI: 10.30595/juita.v10i2.14940
Bi-clustering is a clustering development that aims to group data simultaneously from two directions. The Iterative Signature Algorithm (ISA) is one of the bi-clustering algorithms that work iteratively to find the most correlated bi-cluster. Detecting economic and pandemic vulnerability using bi-cluster analysis is essential to get spatial patterns and an overview of Indonesia's economic and pandemic vulnerability characteristics. Bi-clustering using ISA requires setting the row and column threshold to form seventy combinations of thresholds. The best is chosen based on the average value of mean square residue to volume ratios. In addition, the similarity of the best bi-cluster with the other is also seen based on the Liu and Wang index values. The -1.0 row and -1.0 column threshold combinations were selected and produced the best bi-cluster with the smallest average value of mean square residue to volume ratios (0.00141). Based on Liu and Wang index values, it has more than 95% similarity with the combination of -1.0 row and -0.9 column thresholds and the -0.9 row and -1.0 column thresholds. These selected threshold combinations produce three bi-clusters with five types of spatial patterns and different characteristics because of the overlap between these three bi-clusters.
Penentuan Faktor Kemiskinan Provinsi Banten dengan Model Panel Spasial
Irfani Azis;
I Made Sumertajaya
SAINTIFIK Vol 9 No 1 (2023): Saintifik: Jurnal Matematika, Sains, dan Pembelajarannya
Publisher : Universitas Sulawesi Barat
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.31605/saintifik.v9i1.395
Tujuan dari penelitian ini adalah untuk mengidentifikasi faktor-faktor yang mempengaruhi tingkat kemiskinan kabupaten/kota di Banten dengan menggunakan model data panel spasial. Model data panel yang dipilih adalah model fixed effect, namun model data panel spasial terbaik adalah model SAR yang menggunakan pendekatan matriks ketetanggaan ratu. Model ini memenuhi semua asumsi dan memberikan probabilitas log sebesar 59,83. Menurut model terbaik, variabel yang memiliki dampak terbesar terhadap tingkat kemiskinan adalah jumlah penduduk dan PDRB. Model SAR dan SEM dengan pembobot Queen Contiguity dan KNN menghasilkan dua peubah yang berpengaruh terhadap tingkat kemiskinan Provinsi Banten yaitu PDRB dan jumlah penduduk. Model data panel spasial SEM dengan pengaruh tetap dan pembobot matriks queen contiguity adalah model yang tepat untuk tingkat kemiskinan provinsi Banten karena menghasilkan nilai Loglikelihood yang paling besar.