Tarno Tarno
Departemen Statistika, FSM, Universitas Diponegoro, Jl. Prof Soedharto SH Tembalang, Semarang

Published : 29 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 29 Documents
Search

PERBANDINGAN MODEL JARINGAN SYARAF TIRUAN DENGAN ALGORITMA LEVENBERG-MARQUADT DAN POWELL-BEALE CONJUGATE GRADIENTPADA KECEPATAN ANGIN RATA-RATA DI KOTA SEMARANG Dwi Ispriyanti; Alan Prahutama; Tarno Tarno; Budi Warsito; Hasbi Yasin; Pandu Anggara
Jurnal Statistika Universitas Muhammadiyah Semarang Vol 8, No 2 (2020): Jurnal Statistika Universitas Muhammadiyah Semarang
Publisher : Department Statistics, Faculty Mathematics and Natural Science, UNIMUS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26714/jsunimus.8.2.2020.127-133

Abstract

Wind is one of the most important weather components. Wind is defined as the dynamics of horizontal air mass displacement measured in two parameters, namely speed and direction. Wind speed and direction depend on the air pressure conditions around the place. High wind speed intensity can cause high sea water waves. To estimate wind speed intensity required a study of wind speed prediction. One of method that can be used is Artificial Neural Network (ANN). In ANN there are several models, one of which is backpropagation. Thepurpose of this researchis to compare between backpropagation model with Levenberg-Marquadt and Powell-Beale Conjugate Gradient algorithms. The results of this researchshowing that Powell-Beale Conjugate Gradient better than Levenberg-Marquadtalgorithms. The best model architecture obtained is a network with two input layer neurons, six hidden layer neurons, and one output layer neuron. The activation function used are the logistic sigmoid in the hidden layer and linear in the output layer. MAPE value based on the chosen model is 0,0136% in training process and 0,0088% in testing process.
PENGUKURAN RISIKO GLUE-VALUE-AT-RISK PADA DATA DISTRIBUSI ELLIPTICAL (Studi Kasus: Data Saham PT Indocement Tunggal Prakarsa Tbk, PT Unilever Indonesia Tbk, PT United Tractors Tbk, Periode 1 Juni 2018 – 29 November 2019) Dede Andrianto; Di Asih I Maruddani; Tarno Tarno
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 (762.944 KB) | DOI: 10.14710/j.gauss.v9i1.27525

Abstract

Risk measurement is carried out to determine the risk. Popular methods that can be used to measure risk at a confidence level are Value-at-Risk (VaR) and Tail-Value-at-Risk (TVaR). A Risk measurement should satisfy: translation invariance, positive homogenicity, monocity and subadditivity. VaR does not satisfy one of coherent axioms, namely subadditivity. TVaR is considered capable of overcoming VaR problems, but it’s too large for a risk measure. Glue-Value-at-Risk (GlueVaR) is a method that can overcome these problems because it can be valued between VaR and TVaR and fulfills four coherent axioms. In this paper GlueVaR used in the elliptical distribution for normal distribution to measure the risk of the stock of PT Indocement Tunggal Prakarsa Tbk (INTP), PT Unilever Indonesia Tbk (UNVR), and PT United Tractors Tbk (UNTR) for the period June 1st 2018 – 29th November 2019. After knowing the stock return is normally distributed and used confidence levels of α = 95% and β = 98%, a high selection of distortion ℎ1=0,3≤1−????1−???? and ℎ2=0,4≥ℎ1. The high distortion selected makes GlueVaR worth between VaR and TVaR. GlueVaR for INTP, UNVR, and UNTR respectively are 4.886%; 2.999%; and 4.083%. Thus the lowest risk level is PT Unilever Indonesia Tbk.Keywords : Value-at-Risk, Tail-Value-at-Risk, Glue-Value-at-Risk
PERBANDINGAN METODE DOUBLE EXPONENTIAL SMOOTHING HOLT DAN FUZZY TIME SERIES CHEN UNTUK PERAMALAN HARGA PALADIUM Anes Desduana Selasakmida; Tarno Tarno; Triastuti Wuryandari
Jurnal Gaussian Vol 10, No 3 (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.v10i3.32782

Abstract

Palladium is one of the precious metal commodities with the best performance since 3 years ago. Palladium has many benefits, including being used in the electronics, medical, jewelry and chemical industries. The benefits of palladium in the chemical field are that it can help speed up chemical reactions, filter out toxic gases in exhaust gases, and convert the gas into safer substances, so palladium is usually used as a catalyst for cars. Forecasting is a process of processing past data and projected for future interest using several mathematical models. The model used in this study is the Double Exponential Smoothing Holt and Fuzzy Time Series Chen methods. The process of forecasting palladium prices using monthly data from January 2011 to December 2020 with the Double Exponential Smoothing Holt method and the Fuzzy Time Series Chen method will be carried out in this study to describe the performance of the two methods. Based on the results of the analysis, it can be concluded that the Double Exponential Smoothing Holt and Fuzzy Time Series Chen methods have equally good performance with sMAPE values of 6.21% for Double Exponential Smoothing Holt and 9.554% for Fuzzy Time Series Chen. Forecasting for the next 3 periods using these two methods generally produces forecasting values that are close to the actual data. 
PERBANDINGAN METODE NAÏVE BAYES DAN BAYESIAN REGULARIZATION NEURAL NETWORK (BRNN) UNTUK KLASIFIKASI SINYAL PALSU PADA INDIKATOR STOCHASTIC OSCILLATOR (Studi Kasus: Saham PT Bank Rakyat Indonesia (Persero) Tbk Periode Januari 2017 – Agustus 2019) Fredy Yoseph Marianto; Tarno Tarno; Di Asih I Maruddani
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 (762.553 KB) | DOI: 10.14710/j.gauss.v9i1.27520

Abstract

Keputusan untuk membeli atau menjual saham merupakan kunci utama untuk memperoleh keuntungan dalam trading dan investasi. Salah satu indikator yang dapat digunakan dalam menentukan momentum untuk membeli atau menjual saham adalah Stochastic Oscillator. Sebagai indikator yang sensitif terhadap pergerakan harga saham, Stochastic Oscillator sering mengeluarkan sinyal palsu yang mengakibatkan kerugian dalam transaksi. Terdapat 9 atribut yang diduga dapat mengidentifikasi apakah suatu sinyal yang keluar dari indikator Stochastic Oscillator merupakan sinyal palsu atau tidak. Tujuan dari penelitian ini adalah melakukan klasifikasi atau deteksi sinyal dengan metode Naïve Bayes dan Bayesian Regularization Neural Network (BRNN), dan kemudian membandingkan tingkat akurasi hasil klasifikasi antara kedua metode. Hasil dari penelitian ini menunjukkan bahwa hanya terdapat 6 atribut yang dapat digunakan untuk mengidentifikasi apakah suatu sinyal yang keluar merupakan sinyal palsu atau tidak, yaitu kondisi IHSG, kondisi high price, kondisi low price, kondisi close price, posisi %K, dan posisi %D, serta tingkat akurasi dari metode Naïve Bayes adalah sebesar 76,92%, sedangkan akurasi dari metode BRNN adalah sebesar 80,77%. Dapat disimpulkan bahwa dalam penelitian ini, metode BRNN lebih baik dibandingkan dengan metode Naïve Bayes untuk mendeteksi sinyal palsu yang keluar dari indikator Stochastic Oscillator.Kata kunci: Stochastic Oscillator, Sinyal Palsu, Klasifikasi, Naïve Bayes, BRNN, Akurasi
ANALISIS FAKTOR-FAKTOR YANG MEMPENGARUHI PENERIMA BERAS RASKIN MENGGUNAKAN REGRESI LOGISTIK BINER DENGAN GUI R Agustinus Salomo Parsaulian; Tarno Tarno; Dwi Ispriyanti
Jurnal Gaussian Vol 10, No 1 (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.v10i1.30934

Abstract

The Rice Subsidy Program for Low-Income Communities or the Raskin Program is one of the government's programs to eradicate poverty. However, in practice, determining the criteria for Raskin recipients is a complicated problem. The Raskin program is a cross-sectoral national program both horizontally and vertically, to help meet the rice needs of low-income citizens. Determining the criteria for Raskin recipients is often a complicated issue. This study aims to analyze the classification of the Target Households (RTS) for the Raskin Program. The method used is binary logistic regression by utilizing R GUI. Binary logistic regression method is a method to find the relationship between independent and dependent variables, with a binary or dichotomous dependent variable. The data used is the March 2018 National Socio-Economic Survey (Susenas) data for Brebes Regency. The independent variables used in this study are the criteria for determining poor households, namely the area of the house, floor type of the house, wall type of the house, defecation facilities, lighting used, fuel used, ability to buy meat/milk, education level of the head of the household, and the capacity of installed electricity in the main residence. The results of the analysis show that in the final model, the variables that significantly affect the classification of RTS are the ability to eat healthy food, the capacity of installed electricity in the main residence, the education level of the head of the household, and defecation facilities with an accuracy value of 85.4%.Keywords: Raskin Program, Binary Logistic Regression, R GUI
PEMILIHAN INPUT MODEL ADAPTIVE FUZZY INFERENCE SYSTEM (ANFIS) BERBASIS LAGRANGE MULTIPLIER TEST DILENGKAPI GUI MATLAB (Aplikasi pada Data Harga Beras Kualitas Rendah di Indonesia Periode Januari 2013 – Februari 2019) Khusnul Umi Fatimah; Tarno Tarno; Abdul Hoyyi
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 (773.8 KB) | DOI: 10.14710/j.gauss.v8i4.26725

Abstract

Adaptive Neuro Fuzzy Inference System (ANFIS) is a method that uses artificial neural networks to implement fuzzy inference systems. The optimum ANFIS model is influenced by the selection of inputs, number of membership and rules. In general, the selection of ANFIS input is based on Autoregressive (AR) unit as a result of ARIMA preprocessing. Thus it requires several assumptions. In this research, an alternative selection of ANFIS input based on Lagrange Multiplier Test (LM Test) is used to test hypothesis for the addition of one input. Preprocessing is conducted to obtain the value of partial autocorrelation against Zt. The input lag variable which has the highest partial autocorrelation is the first input ANFIS. The next input selection is selected based on LM test for adding one variable. To test the performance of LM Test, an empirical study of two groups of generated data and low quality rice prices is conducted as a case study. Generating data with stationary and non-stationary criteria has a good performance because it has very good forecasting ability with MAPE out sample for each characteristic are 5.6785% and 9.4547%. In the case study using LM Test, the selected input are and  with the number of membership 2. The chosen model has very good forecasting ability with MAPE outsampel 6.4018%. Keywords : ANFIS, ANFIS Input, LM-Test, Low Quality Rice Prices, Forecasting
EXPECTED SHORTFALL DENGAN PENDEKATAN GLOSTEN-JAGANNATHAN-RUNKLE GARCH DAN GENERALIZED PARETO DISTRIBUTION Lina Tanasya; Di Asih I Maruddani; Tarno Tarno
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.v9i4.29447

Abstract

Stock is a type of investment in financial assets that are many interested by investors. When investing, investors must calculate the expected return on stocks and notice risks that will occur. There are several methods can be used to measure the level of risk one of which is Value at Risk (VaR), but these method often doesn’t fulfill coherence as a risk measure because it doesn’t fulfill the nature of subadditivity. Therefore, the Expected Shortfall (ES) method is used to accommodate these weakness. Stock return data is time series data which has heteroscedasticity and heavy tailed, so time series models used to overcome the problem of heteroscedasticity is GARCH, while the theory for analyzing heavy tailed is Extreme Value Theory (EVT). In this study, there is also a leverage effect so used the asymmetric GARCH model with Glosten-Jagannathan-Runkle GARCH (GJR-GARCH) model and the EVT theory with Generalized Pareto Distribution (GPD) to calculate ES of the stock return from PT. Bank Central Asia Tbk for the period May 1, 2012-January 31, 2020. The best model chosen was ARIMA(1,0,1) GJR-GARCH(1,2). At the 95% confidence level, the risk obtained by investors using a combination of GJR-GARCH and GPD calculations for the next day is 0.7147% exceeding the VaR value of 0.6925%. 
ROBUST SPATIAL AUTOREGRESSIVE UNTUK PEMODELAN ANGKA HARAPAN HIDUP PROVINSI JAWA TIMUR Hidayatul Musyarofah; Hasbi Yasin; Tarno Tarno
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 (1165.264 KB) | DOI: 10.14710/j.gauss.v9i1.27521

Abstract

Spatial regression analysis is regression method used for type of data has a spatial effect. Spatial regression showing the presence of spatial effects on the response variable (Y) is a Spatial Autoregressive (SAR). Outlier often found in research spatial data. The outlier is called the spatial outliers. The analysis can be used to handle outliers in general is Robust Regression. There are several estimator that can be used in which the estimator Robust Regression S, M, MM and LTS. Meanwhile, Robust Regression were used to handle spatial outlier is a combination of SAR and Regression Robust method to form a new method that is Robust Spatial Autoregressive (Robust SAR). Type estimator used in this study is the S-Estimator. This study was conducted to determine the best model on a case study Life Expectancy of East Java Province. The best model is analyzed by comparing the methods of SAR and SAR Robust method. Based on the analysis results obtained MSE and Adjusted R2 values for the SAR method are 1.7521 and 55.54% while for the Robust SAR method are 0.7456 and 62.30%. The Robust SAR model has a lower MSE value and a higher Adjusted R2 when compared to the SAR model. Thus the best model for modeling the life expectancy in East Java is Robust SAR models.Keywords:Spatial Autoregressive (SAR), Robust SAR, Life expectancy
PERAMALAN DATA INDEKS HARGA KONSUMEN KOTA PURWOKERTO MENGGUNAKAN MODEL FUNGSI TRANSFER MULTI INPUT Inarotul Amani Rizki Ananda; Tarno Tarno; Sudarno Sudarno
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.v9i4.29406

Abstract

The Consumer Price Index (CPI) provides information on changes in the average price of a group of fixed goods or services that are generally consumed by households within a certain period of time. The General CPI is formed from 7 sectors of public consumption expenditure groups. Because the formation of the consumer price index value is influenced by several sectors, the method that can be used is the transfer function method. The purpose of this study is to analyze the transfer function model so that the best model is produced to predict CPI in Purwokerto for the next several periods. In this study, general CPI modeling will be carried out based on the CPI value for the transportation services sector and the CPI for the Health sector in Purwokerto from January 2014 to July 2019 using the multi-input transfer function method. Based on the analysis, the best models are obtained, namely the multi-input transfer function model (2,0,0) (0,1,0) and the ARIMA noise series ([3], 0,0). The model has an Akaike's Information Criterion (AIC) value of 72.42021 and an sMAPE value of  2,351591 % which indicates that the model can be used for forecasting..Keywords: Consumer Price Index (CPI), Inflation,transfer function, AIC
PREDIKSI HARGA JUAL KAKAO DENGAN METODE LONG SHORT-TERM MEMORY MENGGUNAKAN METODE OPTIMASI ROOT MEAN SQUARE PROPAGATION DAN ADAPTIVE MOMENT ESTIMATION DILENGKAPI GUI RSHINY Yayan Setiawan; Tarno Tarno; Puspita Kartikasari
Jurnal Gaussian Vol 11, No 1 (2022): 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.v11i1.33994

Abstract

Cocoa is a leading commodity from Indonesia. Cocoa prices from time to time fluctuate. Accurate Cocoa price predictions are very important to ensure future prices and help decision making. Cocoa price data is non-stationary and nonlinear, so to make accurate predictions, an Artificial Neural Network (ANN) model is applied. One type of ANN is Long Short-Term Memory (LSTM). LSTM has superior performance for time series based prediction. Optimization methods used are Root Mean Square Propagation, and Adaptive Moment Estimation. The best model was selected based on the Means Square Error (MSE) and Mean Absolute Percentage Error (MAPE) values. This study uses the R-Shiny GUI to facilitate the use of LSTM for users who are less proficient in programming languages. Based on the results, the Long Short-Term Memory model with the Adaptive Moment Estimation optimization method is more optimal than the Long Short-Term Memory with Root Mean Square Propagation seen from the smaller MSE and MAPE values. This study used 27 combinations of hyperparameters. Prediction results with LSTM using the R-Shiny GUI have different levels of accuracy in each experiment. The best accuracy value is experiment with MSE value of 491505.1 and MAPE value of 1.739155% . Cocoa Price Forecasting for the period November to December 2021 tends to decline.Keywords : Cocoa Prices, Forecasting, Long Short-Term Memory, Root Mean Square Propagation, Adaptive Moment Estimation, GUI R-Shiny