Claim Missing Document
Check
Articles

PERAMALAN JUMLAH KUNJUNGAN WISATAWAN MANCANEGARA DI KEPULAUAN RIAU DENGAN MENGGUNAKAN MODEL FUNGSI TRANSFER Tamura Rolasnirohatta Siahaan; Rukun Santoso; Alan Prahutama
Jurnal Gaussian Vol 9, No 2 (2020): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (513.88 KB) | DOI: 10.14710/j.gauss.v9i2.27817

Abstract

Transfer function models is a data analysis model that combines time series and causal approach, in another words, transfer function models is a method that ilustrates that the predicted value in teh future is affected by the past value time series and based on one or more related time series. In this research, an analysis of the number of tourist arrival and rainfall in several regions in Kepulauan Riau from January 2013 until December 2017 was aimed at obtaining a transfer function model and forecasting the number of tourist arrival in several regions of the Kepulauan Riau for next periods. Based on the result of the analysis, rainfall in Tanjung Pinang does not affect the visit of tourist with the values of MAPE is 13,63494%. Rainfall in Batam also does not affect the visit of tourist with the values of MAPE is 7,977151%. While in Tanjung Balai Karimun, tourist arrivals was affected by rainfall with the values of MAPE is 10,32777%.
PEMODELAN GEOGRAPHICALLY WEIGHTED GENERALIZED POISSON REGRESSION (GWGPR) PADA KASUS KEMATIAN IBU NIFAS DI JAWA TENGAH Wahyu Sabtika; Alan Prahutama; Hasbi Yasin
Jurnal Gaussian Vol 10, No 2 (2021): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/j.gauss.v10i2.30946

Abstract

Maternal mortality is one indicator to describing prosperity in a country and indicator of women's health. Most of the maternal mortality caused by postpartum maternal mortality. The number of postpastum maternal mortality is events that the probability of the incident is small, where the incident depending on a certain time or in a certain regions with the results of the observation are variable diskrit and between variable independent each other that follows the Poisson distribution, so that the proper statistical method is Poisson regression. However, in Poisson regression model analysis sometimes assumptions can occur violations, where the value of variance is greater than the mean value called overdispersion. Generalized Poisson Regression (GPR) is one model that can be used to handle overdispersion problems. This modeling produces global parameters for all locations (regions), so to overcome this we need a method of statistical modeling with due regard to spatial factors. The analytical method used to determine the factors that influence the number of postpartum maternal mortality in Central Java that have overdispersion and there are spatial factors, is Geographically Weighted Generalized Poisson Regression (GWGPR) using the Maximum Likelihood Estimation method and Adaptive Bisquare weighting. Poisson regression and GPR modeling produces a variable percentage of pregnant women doing K1 which has a significant effect on the number of postpartum maternal mortality, while for GWGPR modeling is divided into four cluster in all regency/city in Central Java based on the same significant variable. From the comparison of AIC values, it was found that the GWGPR model is better for analyzing postpartum maternal mortality in Central Java because it has the smallest AIC value.Keywords: The Number of Postpartum Maternal Mortality, Overdispersion, Generalized Poisson Regression, Spatial, Geograpically Weighted Generalized Poisson Regression, AIC
PERBANDINGAN METODE MOORA DAN TOPSIS DALAM PENENTUAN PENERIMAAN SISWA BARU DENGAN PEMBOBOTAN ROC MENGGUNAKAN GUI MATLAB Rafida Zahro Hasibuan; Alan Prahutama; Dwi Ispriyanti
Jurnal Gaussian Vol 8, No 4 (2019): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (881.296 KB) | DOI: 10.14710/j.gauss.v8i4.26726

Abstract

MAN Asahan is an educational institution that selects new students every year. MAN Asahan sets certain criteria in choosing new students so that selected students are of high quality. The criteria determined are the Al-Qur'an test scores, national exam scores, Academic Potential Test scores and achievement certificates. In selecting new students who were accepted as many as 271 of the 530 registrants the school still used the manual process so that it needed accuracy and a long time. In this study a decision support system was created that could be a solution to assist the selection process according to school criteria. The system will applied is MOORA (Multi-Objective Optimization on the Base of Ratio Analysis) method and TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) with the weighting method of ROC (Rank Order Centroid). Then the sensitivity analysis is done to determine the appropriate method to be chosen to obtain optimal results. This research was conducted with the help of the MATLAB GUI as a computing tool. The GUI that is built can simplify and speed up the selection process. Based on the results of the study, the average percentage value of sensitivity for the MOORA method is -1.61% while the TOPSIS method is -7.96%. With the existence of sensitivity analysis it can be known the most appropriate method for this case is the MOORA method.Keywords: Students, MOORA, TOPSIS, ROC, Sensitivity, GUI Matlab
PERBANDINGAN NILAI KORELASI PADA KANONIK ROBUST (METODE MINIMUM COVARIANCE DETERMINANT) DAN KANONIK KLASIK (Studi Kasus Data Struktur Ekonomi dan Kesejahteraan Rakyat di Jawa Barat 2016) Widi Rahayu; Sudarno Sudarno; Alan Prahutama
Jurnal Gaussian Vol 8, No 4 (2019): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (795.636 KB) | DOI: 10.14710/j.gauss.v8i4.26753

Abstract

Canonical correlation analysis is a multivariate statistical analysis that aims to examine the correlation between two groups of variabels in a way to maximize the value of correlation between variabels. The outlier in the data affect the covariance matrix is generated, So that use robust multivarat. There is robust multivariate approach to the analysis of canonical robust with MCD method (Minimum Covariance Determinant). This final project aims to determine comparison between correlation value of robust canonical with MCD and canonical classical methods. With a data theres containing of outliers in the case studies of people's welfare and economic structures in West Java in 2016. Used a set of variabels welfare of people consist of 6 variabel (Y) and a set of variabels economic structure which consists of four variabels (X). Based on the analysis results obtained that robust canonical correlation values better explain the correlation between two sets of variabels, the correlation value 0.99552, =0.91228, =0.71529, =0.63174, While the correlation value on classical canonical are 0.931489, 0.538672, 0.387099, 0.259318, Canonical robust can be interpreted more because it meets the test of significance are partially and directly, while the classical canon can not be interpreted further because it does not meet the test of the significance of the function. Keywords       : Classical canonical correlation, canonical correlation robust correlation value, Minimum Covariance Determinant (MCD)
PEMODELAN KECEPATAN ANGIN DI KOTA SEMARANG MENGGUNAKAN ADAPTIVE NEURO-FUZZY INFERENCE SYSTEM (ANFIS) Alifah Zahlevi; Alan Prahutama; Abdul Hoyyi
Jurnal Gaussian Vol 8, No 3 (2019): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (493.913 KB) | DOI: 10.14710/j.gauss.v8i3.26709

Abstract

Semarang city is the one of the strategic areas located in the middle of the north coast of Java that has a tropical climate with the high humidity and temperature, so it often causes a high rainfall and strong wind. So that is way Semarang city is ever sustained the extreme weather like a Tropical Storm. Since January 2016 until 2017 there are 34 cases of Tornado and 24 incidents of fallen trees because of the gale. For helping the people to be allert the effect of the strong winds can be done by predicting the average of wind velocity by using Adaptive Neuro-Fuzzy Inference System (ANFIS) method which can predict the climate change that do not require the assumption of white noise and normal residual distribution. In addition ANFIS is a group of neural network with input that has been fuzzied on the first or second layer, but the weight of the artificial neural is not fuzzied. The identification result of stationaries obtained the plot of PACF on the first and second lag, with the result that these lag which will be a input variable on ANFIS model. The result of ANFIS by using cluster FCM, the third total membership show the smallest percentage of RMSE in-sample is 0,0048 on the first lag, and the smallest percentage of RMSE out-sample is 0,008 on the ANFIS model with the input lag 1 and three cluster. Keywords : the average of wind velocity, ANFIS, RMSE
ESTIMASI VALUE AT RISK PORTOFOLIO SAHAM MENGGUNAKAN METODE GARCH-COPULA (Studi Kasus : Harga Penutupan Saham Harian Unilever Indonesia dan Kimia Farma Periode 1 Januari 2013- 31 Desember 2016) Lingga Bayu Prasetya; Dwi Ispriyanti; Alan Prahutama
Jurnal Gaussian Vol 7, No 4 (2018): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/j.gauss.v7i4.28867

Abstract

Any investment in the stock market will earn returns accompanied by risks. Return and risk has a mutual correlation that equilibrium. The formation of a portfolio is intended to provide a lower risk or with the same risk but provide a higher return. Value at Risk (VaR) is a instrument to analyze risk management. Time series model used in stock return data that it has not normal distribution and heteroscedastisicity is Generalized Autoregressive Conditional Heteroscedasticity (GARCH). GARCH-Copula is a combined method of GARCH and Copula. The Copula method is used in joint distribution modeling because it does not require the assumption of normality of the data and can capture tail dependence between each variable. This research uses return data from stock closing prices of Unilever Indonesia and Kimia Farma period January 1, 2013 until December 31, 2016. Copula model is selected based on the highest likelihood log value is Copula Clayton. Value at Risk estimates of Unilever Indonesia and Kimia Farma's stock portfolio on the same weight were performed using Monte Carlo simulation with backtesting of 30 days period data at 95% confidence level. Keywords : Stock, Risk, Generalized Autoregressive Conditional Heteroscedasticity (GARCH), Copula, Value at Risk
KLASIFIKASI STATUS KEMISKINAN RUMAH TANGGA DENGAN METODE SUPPORT VECTOR MACHINES (SVM) DAN CLASSIFICATION AND REGRESSION TREES (CART) MENGGUNAKAN GUI R (Studi Kasus di Kabupaten Wonosobo Tahun 2018) Lutfia Nuzula; Alan Prahutama; Arief Rachman Hakim
Jurnal Gaussian Vol 9, No 4 (2020): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/j.gauss.v9i4.29449

Abstract

The poor are people who have average monthly expenditures per capita below the poverty line. Wonosobo District became the poorest district in Central Java in 2011-2018, although the percentage of poor people has decreased every year. It cannot be separated from the efforts of the Wonosobo District Government to overcome poverty through various programs. This study classified households in Wonosobo District in 2018 as poor and non-poor based on influencing factors. This study used the Support Vector Machines (SVM) method to be compared with the Classification and Regression Trees (CART) method. It used the data from the 2018 National Socio-Economic Survey of Central Java with a total of 795 observations. Result of the research using the SVM method and the RBF kernel, the classification accuracy reaches 89.82% then the classification accuracy using the CART method reaches 87.08%. GUI designed by RShiny package can make easier for users to analyze the SVM and CART with the valid output. 
PEMODELAN REGRESI SEMIPARAMETRIK DENGAN PENDEKATAN DERET FOURIER (Studi Kasus: Pengaruh Indeks Dow Jones dan BI Rate Terhadap Indeks Harga Saham Gabungan Laili Rahma Khairunnisa; Alan Prahutama; Rukun Santoso
Jurnal Gaussian Vol 9, No 1 (2020): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (791.883 KB) | DOI: 10.14710/j.gauss.v9i1.27523

Abstract

The Composite Stock Price Index (CSPI) is a composite index all of types of shares listed on the stock exchange and their movements indicate conditions that occur in the capital market. CSPI is influenced by macroeconomic factors and foreign exchange index. Dow Jones Industrial Average has a linear relationship with CSPI and BI Rate has a repeated relationship with CSPI, so the method is used semiparametric regression with the Fourier series approach. Estimators in semiparametric regression with Fourier series approach were obtained by the Ordinary Least Square (OLS) method. This study uses monthly data which is divided into in sample data and out sample data. Semiparametric regression modelling with Fourier series approach is done by determining the optimal K value which results in a minimum General Cross Validation (GCV) value. In this study, semiparametric regression model with Fourier series approach formed by the optimal K value is 13 and GCV is 2826122. The results of the evaluation of the accuracy of the model performance and forecasting obtained the coefficient of determination is 0,9226, Mean Absolute Percentage Error (MAPE) data in sample 3,8154% and data out sample is 8,4782% which shows that the model obtained has a very accurate performance.Keywords: Composite Stock Price Index (CSPI), Semiparametric Regression, Fourier Series, OLS, GCV
PERAMALAN JUMLAH PENUMPANG KERETA API MENGGUNAKAN METODE ARIMA, INTERVENSI DAN ARFIMA (Studi Kasus : Penumpang Kereta Api Kelas Lokal EkonomiDAOP IV Semarang) Helmi Panjaitan; Alan Prahutama; Sudarno Sudarno
Jurnal Gaussian Vol 7, No 1 (2018): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (607.933 KB) | DOI: 10.14710/j.gauss.v7i1.26639

Abstract

Autoregressive Integrated Moving Average (ARIMA) is stationary time series model after differentiation. Differentiation value of ARIMA method is an integer so it is only able to model in the short term. The best model using ARIMA method is ARIMA([13]; 1; 0) with an MSE value of 1,870844. The Intervention method is a model for time series data which in practice has extreme fluctuations both up and down. In the data plot the number of train passengers was found to be extreme fluctuation. The data used was from January 2009 to June 2017 where fluctuation up significantly in January 2016 (T=85 to T=102) so the intervention model that was suspected was a step function. The best model uses the Intervention step function is ARIMA ([13]; 1; 1) (b=0; s=18; r=0) with MSE of 1124. Autoregressive Fractionally Integrated Moving Average (ARFIMA) method is a development of the ARIMA method. The advantage of the ARFIMA method is the non-integer differentiation value so that it can overcome long memory effect that can not be solve with the ARIMA method. ARFIMA model is capable of modeling high changes in the long term (long term persistence) and explain long-term and short-term correlation structures at the same time. The number of local economy class train passengers in DAOP IV Semarang contains long memory effects, so the ARFIMA method is used to obtain the best model. The best model obtained is the ARMA(0; [1,13]) model with the differential value is 0,367546, then the model can be written into ARFIMA (0; d; [1,13]) with an MSE value of 0,00964. Based on the analysis of the three methods, the best method of analyzing the number of local economy class train passengers in DAOP IV Semarang is the ARFIMA method with the model is ARFIMA (0; 0,367546; [1,13]). Keywords: Train Passengers, ARIMA, Intervention, ARFIMA, Forecasting
IMPLEMENTASI SUBSET AUTOREGRESSIVE MENGGUNAKAN PAKET FITAR Tomi Ardi; Rukun Santoso; Alan Prahutama
Jurnal Gaussian Vol 6, No 4 (2017): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/j.gauss.v6i4.30385

Abstract

Time series data analysis is one of the important points in statistics that is a time-dependent analysis. The commonly used model for time series data is ARIMA (Autoregressive Integrated Moving Average) or often also called the Box-Jenkins time series method. A model of ARIMA used in time clock data forecasting is the AR subset (autoregressive). The AR subset model is suitable for a long time series with a more than 5th order lag. The statistical software used is the R. time series AR subset approach on R using the FitAR package. The main function of the FitAR package is SelectModel and FitAR. SelectModel function to get the model automatically while FitAR is used to determine the temporary suspect model. Data used in the form of dataset contained in package FitAR that is SeriesA. The SeriesA data is data about the chemical concentration process observed every 2 hours for 17 days. SeriesA is processed using FitAR package so that the best model is AR [1,2,7].Keywords : Time Series, Time Series Non-stasioner, Subset AR, FitAR Package