Papers by Keyword: Time Series Analysis

Abstract: Traditionally the chemical industry uses coal, minerals and petroleum as its basic raw materials, but palm oil and palm kernel oil are being increasingly used as economical raw materials especially for the production of oleochemicals. High magnitude palm oil price volatility in recently years has been a major challenge faced the industry. Though many time series models have been developed, few have wide adoption in the industry, and one of the key issues is the sampling interval used in the models. To date, little effort has been spent on mining historical data to determine the representativeness of interval sampling. This paper presents a novel approach in identifying price equilibrium for crude palm oil by mining the sampling amount through historical price distribution. Evaluation is done on the outcomes of the experiment, and analysis is performed on the attributes of each different criteria of the price distribution. The performance of the proposed approach is also compared to the conventional Bollinger Band with static sampling size. Overall, the preliminary results show that price distribution with leptokurtic distribution outperforms other price distribution patterns, this will definitely assist further works to devise a novel financial time series analysis technique.

Abstract: The monitoring of processes is a vital mechanism for ensuring that such processes remain safe and under control. The present research aims to solve problems associated with correlated data by applying the Box-Jenkins method integrated with statistical process control (SPC) tools, namely the Shewhart chart, the moving average chart, the cumulative sum (CUSUM) chart, and the exponentially weighted moving-average (EWMA) chart. The efficiency of the four SPC tools was also compared in terms of the false alarm rate (FAR) and the missed detection rate (MDR). The findings indicated that the EWMA chart was the most effective in detecting anomaly, the Shewhart chart and the moving average chart produced high MDR, and the CUSUM chart suffered the highest FAR.

Abstract: The paper provides analysis of new car sales in Bosnia and Herzegovina for the period 2007-2014. For the examined period new car sales in B&H was reduced by 53.67%, from 12449 in 2007 to 6682 in 2014. The trend can be approximated using 3rd degree polynomial regression model with coefficient of determination R²=0,845. Most new cars sold were by Skoda and least was sold by Porsche. Total number of sold vehicles for this period was 73152. We also present annual growth, chain growth and cumulative growth index for the given period.

Abstract: Ethanol is the important renewable energy in Thailand. It is alcohol produced from sugarcane and tapioca that are agricultural products available in Thailand. Ethanol is used to blend with gasoline for use as gasohol. Ethanol production and consumption in Thailand are fluctuating. Consequently, planning of ethanol production and consumption is irrelevant. In order to solve this problem, this study aims to find forecasting models using time series analysis including exponential smoothing and the Box-Jenkins methods. The most appropriate forecasting model was selected from the two methods by considering the minimum of the mean absolute percentage error: MAPE. It was found that the Box-Jenkins is the most appropriate method to forecast both ethanol production and consumption. The forecasting results were then used to determine appropriate quantity and proportion of molasses and tapioca needed for ethanol production in the future.

Abstract: The amount of municipal solid waste(MSW)transportation has become an important basis of handling urban domestic waste, at the same time, accurate predictions of time series data have motivated the researchers to develop innovative models for urban solid waste management.Therefore, predicting the MSW transportation amount in a scientific manner is one of the most essential parts of the urban waste management work. Based on the raw data of MSW transportation amount from 1993 to 2012 of Wuhan city, the capital of Hubei province, this paper chose Autoregressive Integrated Moving Average Model(also called ARIMA model), used Eviews software to process the data and test various effective inspection, then made a prediction of the amount of MSW transportation of Wuhan, and got access to the conclusions through comparing the original data and predicted one. The results showed that the predicted value of the amount of MSW transportation in 2013 was consistent with the original one, and would reach 214.82 wt in 2014. The results also demonstrated that, the MSW transportation amount prediction based on ARIMA model is practicable due to its high applicability and accuracy, offering decisive information for the urban environmental planning and urban domestic waste controlling.

Abstract: Two new algorithms for nonlinear damage detection are proposed based on linear model with autoregressive moving average (ARMA) in this paper. Firstly, a novel DSF is defined and the DSFs are identified and classified followed by cluster analysis or Bayesian discrimination. Secondly, the performances of the presented algorithms are evaluated and verified by the experimental data of a three-story building structure. Finally, the illustrated results show the algorithms are efficient tools for nonlinear damage detection. They grant a higher accuracy and improve the reliability of nonlinear damage detection whilst reducing computational costs. It can thus be inferred that the proposed algorithms are applicable for Structural Health Monitoring (SHM) in situ.

Abstract: This research presents a time series analysis and artificial neural network (ANN)-based scheme for fault diagnosis of power transformers, which extracts the characteristic parameters of the faults of the transformer from the results of time series analysis and bases on this basis establishes the corresponding back propagation (BP) neural network to detect the transformer operating faults. The simulation experimental results show that as compared to the related works, the proposed approach effectively integrates the superiority of time series analysis and BP neural network and thus can greatly improve the diagnosis accuracy and reliability.

Abstract: High calculation precision and speed of the model parameter estimation has become the theoretical research emphasis and the key link of the applications of the time series analysis based methods. Aiming at the problem that some of the previous parameter estimation methods exist the weakness of stronger constraints, higher time complexity, lower precision of the whole recurrence process and insufficient online diagnosis power, this paper proposes an approach which repeatedly uses the auto-covariance function and the autocorrelation function throughout the recurrent process while guaranteeing not to increase the time complexity of the proposed algorithm and, hence improve the calculation speed and accuracy of parameter estimation simultaneously. As compared to related work, it has lower time complexity, shorter computation time and higher parameter estimation accuracy. The fault diagnosis steps based on the proposed parameter estimation approach are also suggested.

Abstract: Time series analysis is advantageous since it offers insight into the underlying dynamics and forecasts system behavior. The construction of the discriminant function is of vital importance in the time series analysis based fault diagnosis. Aiming at the problem that some of the time series analysis based fault diagnosis methods exist the weakness of higher time complexity, weaker discriminant ability and insufficient online diagnosis power, this paper proposes an approach which makes full use of the characteristics of the model and observation data to construct the discriminant function, and presents an efficient algorithm which can effectively recognize the system state by the proposed discriminant function. As compared to the related work, it has the characteristics of lower time complexity, shorter computation time and stronger distinguished ability, without the requirement of same orders of the reference model and the detected model. The fault diagnosis steps based on the proposed discriminant function and its algorithm are also suggested.

Abstract: In order to forecast the malfunctional state of main power plant on ships through predicting the trend of unsteadily changed thermal performance parameters in main power plant system, this paper established AR model for thermal performance parameters using time series analysis method, by which to predict the change trend analysis of thermal performance parameters, and comparing the predicted values and the measured values. The actual application case of predicting a main engine’s exhaust temperature has been validated by using of the AR model, and results showed that: AR model can accurately predict the change trend of smoke temperature, and the prediction precision is higher, the average relative error was 0.25%.