Forecasting Determinant of Cement Demand in Indonesia with Artificial Neural Network

Fradinata, Edy; Suthummanon, Sakesun; Suntiamorntut, W.

doi:10.18488/journal.2/2015.5.7/2.7.373.384

Cited by 7 publications

(6 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The variables determinatof demand are GDP growth (D1); Population (D2), A potential customer (D3),; Price (D4); Sales (D5); Advertising (D6); Quality (D7); Expectation future price (D8); Preference price,(Trend seasonal) (D9) [9]. The fluctuations of data show the characteristic of time series data set in monthly basis or in 8 years cement demand [10].…”

Section: Results and Analysis 31 Data Experimentsmentioning

confidence: 99%

Initial Optimal Parameters of Artificial Neural Network and Support Vector Regression

Fradinata¹,

Suthummanon²,

Suntiamorntut³

2018

IJECE

Self Cite

View full text Add to dashboard Cite

This paper presents architecture of backpropagation Artificial Neural Network (ANN) and Support Vector Regression (SVR) models in supervised learning process for cement demand dataset. This study aims to identify the effectiveness of each parameter of mean square error (MSE) indicators for time series dataset. The study varies different random sample in each demand parameter in the network of ANN and support vector function as well. The variations of percent datasets from activation function, learning rate of sigmoid and purelin, hidden layer, neurons, and training function should be applied for ANN. Furthermore, SVR is varied in kernel function, lost function and insensitivity to obtain the best result from its simulation. The best results of this study for ANN activation function is Sigmoid. The amount of data input is 100% or 96 of data, 150 learning rates, one hidden layer, trinlm training function, 15 neurons and 3 total layers. The best results for SVR are six variables that run in optimal condition, kernel function is linear, loss function is ౬-insensitive, and insensitivity was 1. The better results for both methods are six variables. The contribution of this study is to obtain the optimal parameters for specific variables of ANN and SVR.

show abstract

Section: Results and Analysis 31 Data Experimentsmentioning

confidence: 99%

Initial Optimal Parameters of Artificial Neural Network and Support Vector Regression

Fradinata¹,

Suthummanon²,

Suntiamorntut³

2018

IJECE

Self Cite

View full text Add to dashboard Cite

show abstract

“…ANN has been implemented in many areas for forecasting method to obtain the best result, due to its flexibility since the ANN produce the good prediction Compare the forecasting method of artificial neural network and support vector regression model to measure the bullwhip effect in supply chain 4820 model [38,39]. It does not require many statistical assumptions, continue and non-continue, parametric and nonparametric data; it can also work in missing dataset [40]. ANN prediction used the back-propagation Levenberg-Marquardt training algorithm that contributes to the good result.…”

Section: Methodsmentioning

confidence: 99%

Compare the forecasting method of artificial neural network and support vector regression model to measure the bullwhip effect in supply chain

Fradinata

Suthummanon

Suntiamorntut

et al. 2019

JMES

View full text Add to dashboard Cite

The objective of this study was to compare the Bullwhip Effect (BWE) in the supply chain through two methods and to determine the inventory policy for the uncertainty demand. It would be useful to determine the best forecasting method to predict the certain condition. The two methods are Artificial Neural Network (ANN) and Support Vector Regression (SVR), which would be applied in this study. The data was obtained from the instant noodle dataset where it was in random normal distribution. The forecasting demands signal have Mean Squared Error (MSE) where it is used to measure the bullwhip effect in the supply chain member. The magnification of order among the member of the supply chain would influence the inventory. It is quite important to understand forecasting techniques and the bullwhip effect for the warehouse manager to manage the inventory in the warehouse, especially in probabilistic demand of the customer. This process determines the appropriate inventory policy for the retailer. The result from this study shows that ANN and SVR have the variance of 0.00491 and 0.07703, the MSE was 1.55e-6 and 1.53e-2, and the total BWE was 95.61 and 1237.19 respectively. It concluded that the ANN has a smaller variance than SVR, therefore, the ANN has a better performance than SVR, and the ANN has smaller BWE than SVR. At last, the inventory policy was determined with the continuous review policy for the uncertainty demand in the supply chain member.

show abstract

“…The products can be produced based on a certain customer need by the factory. The factory should be on a contract with the retailer where it must send a certain quantity on a specific date based on the agreement (Fradinata et al, 2015).…”

Section: Time-varying Demandmentioning

confidence: 99%

Optimal Lot-Sizing Algorithms on Stochastic Demand at the Retailer

Fradinata¹,

MarliKesuma²,

SitiRusdiana³

2018

Proceedings of MICoMS 2017

Self Cite

View full text Add to dashboard Cite

Purpose-The purpose of this study is to explore the concept of the economic lot sizing and the time cycle period of reordering. The stochastic demand is quite common in the real environment of a cement retailer. The study compares three methods to obtain the optimal solution of a lot-sizing ordering from the real case of the previous study where the dataset is collected from the area of some retailers at Banda Aceh Province of Indonesia. Design/Methodology/Approach-The problem model appears when the retailer with shortage has to fulfill the lot size in the optimal condition to the stochastic demand while at the same time has the backlog condition. Moreover, when the backorder needs the time horizon for replenishment where this condition influences the holding cost at the store, many retailers try to solve this problem to minimize the holding cost, but on the other side, it should fulfill the customer demand. Three methods are explored to identify that condition: a Wagner-Whitin algorithm, the Silver-Meal heuristic, and the holding and ordering costs. The three methods are applied to the lot sizing when there is a backlog. Findings-The results of this study show that the Wagner-Whitin algorithm outperforms the other two methods. It shows that the performance increases around 27% when compared to the two other methods in this study. Research Limitations/Implications-All models are almost approximate and useful to determine the cycle period on stochastic demand. Practical Implications-The calculation of the dataset with the three methods would give the simple example to the retailer when he faces the uncertainty demand models. The prediction of the calculation is done accurately than the constant calculation, which is more economic. Social Implications-The calculation will contribute to much better predictions in many cases of uncertainty.

show abstract

Forecasting Determinant of Cement Demand in Indonesia with Artificial Neural Network

Cited by 7 publications

References 19 publications

Initial Optimal Parameters of Artificial Neural Network and Support Vector Regression

Initial Optimal Parameters of Artificial Neural Network and Support Vector Regression

Compare the forecasting method of artificial neural network and support vector regression model to measure the bullwhip effect in supply chain

Optimal Lot-Sizing Algorithms on Stochastic Demand at the Retailer

Contact Info

Product

Resources

About