模糊時間數列之FSARI(p, d)×(P, D)(下标 s)模型的研究

Many statistical models have been implemented in the energy sectors, especially in the oil production and oil consumption. However, these models required some assumptions regarding the data size and the normality of data set. These assumptions give impact to the forecasting accuracy. In this paper, the fuzzy time series (FTS) model is suggested to solve both problems, with no assumption be considered. The forecasting accuracy is improved through modification of the interval numbers of data set. The yearly oil production and oil consumption of Malaysia from 1965 to 2012 are examined in evaluating the performance of FTS and regression time series (RTS) models, respectively. The result indicates that FTS model is better than RTS model in terms of the forecasting accuracy.

The one central problem in global forecasting area is to minimize the forecasting error and to have a simple model. This paper focuses on the fuzzy time series (FTS) forecasting which is one branch of fuzzy set (FS). Numerous FTS models have been proposed in literature during the past decades and the number of new models is still growing especially in stock markets and student enrolment problems. Some of the FTS models found in literature do not explain the fuzzy logical relationship strength between Ai and Aj which are the left-hand side (LHS) and right-hand side (RHS) of the fuzzy logical relationship respectively. This paper proposed Weighted Subsethood Fuzzy Time Series (WeSuFTS) to build fuzzy reasoning based to address the drawback in forecasting enrolment of students. In this paper, we create the weighted subsethood fuzzy rules to explain the fuzzy relationship between Ai and Aj . The simple centre mid-point defuzzification technique was introduced in defuzzification step. The results show that WeSuFTS method gives promising results where the average error forecasting is 2.31% which is smaller than the one presented by Song and Chissom in 1993 and Chen in 1996, the pioneer of FTS. The proposed method has an advantage since it can give a new interpretation of FTS model. Furthermore, this method able to present the degree of importance of LHS to RHS in explaining the fuzzy logical relationship in FTS.

Non-stationary time series plays a prominent role in the analysis of performance time series of many real-world systems. Recently, fuzzy time series models have been extended to forecast non-stationary time series. Over different phases of time, performance time series may show drastic changes. Therefore, a non-stationary time series is partitioned according to different phases of time. These phases may be taken as weeks, months, or years. Over different phases of time, the universe of discourse, knowledge base, and rule base may vary. A common constraint in the modelling of time-variant fuzzy time series is their incapability to address the phase change of time in the model and accordingly incorporate the necessary changes in the universe of discourse, knowledge base, and rule base over different time phases. To address this issue, a Phase-cum-Time Variant Fuzzy Time Series Model (PTVFTS) is presented in this paper. The PTVFTS model is developed so that it will address the problem of phase change as well as the time variations within each phase simultaneously. The concept of the model rebuilding process is applied to handle changes over each phase and the modified parameter adaptation technique is used for the time variations within each phase. The developed model is applied to the daily closing price time series for the years 2017, 2018, 2019, 2020, 2021, and 2022 separately of stock market indices, NASDAQ, S&P 500, Dow Jones, and TAIEX. The comparison of the developed model is made with the time-variant fuzzy time series model known as the non-stationary fuzzy time series model (NSFTS), Dynamic Evolving Neural-Fuzzy Inference System (DENFIS) model, Long Short-Term Memory (LSTM) model, and the classical Auto-Regressive Integrated Moving Average (ARIMA) model. The efficiency of the developed PTVFTS model is tested using forecasting metrics and statistical tests. The comparison shows that the developed model is more efficient in forecasting phase-cum-time variant non-stationary time series.

Fuzzy time series models that have been developed have been widely applied to many applications of forecasting future stock prices or weighted indexes in the financial field. Three interesting problems have been identified in relation to the associated time series methods, as follows: (1) conventional time series models that consider single variables on associated problems only, (2) fuzzy time series models that determine the interval length of the linguistic values subjectively, and (3) selected variables that depend on personal experience and opinion subjectively. In light of the above limitations, this study constitutes a hybrid seven-step procedure that proposes three integrated fuzzy time series models that are based on fitting functions to forecast weighted indexes of the stock market. First, the proposed models employ Pearson correlation coefficients to objectively select important technical indicators. Second, this study utilizes an objective algorithm to determine the lower bound and upper bound of the universe of discourse automatically. Third, the proposed models use the spread-partition algorithm to automatically determine linguistic intervals. Finally, they combine the transformed variables to build three fuzzy time series models using the criterion of the minimal root mean square error (RMSE). Furthermore, this study provides all of the necessary justifying information for using a linear process to select the inputs for the given non-linear data. To further evaluate the performance of the proposed models, the transaction records of TAIEX (Taiwan Stock Exchange Capitalization Weighted Stock Index) and HSI (Hang Seng Indexes) from 1998/01/03 to 2006/12/31 are used to illustrate the methodology with two experimental data sets. Chen's (Fuzzy Sets Syst. 81:311---319, 1996) model, Yu's (Physica A 349:609---624, 2005) model, support vector regression (SVR), and partial least square regression (PLSR) are used as models to be compared with the proposed model when given the same data sets. The analytical results show that the proposed models outperform the listed models under the evaluation criteria of the RMSE (in contrast to the forecasting accuracy) for forecasting a weighted stock index in both the Taiwan and Hong Kong stock markets.

The arising air pollution has addressed much attention globally due to its detrimental effects on human health and environment. As an early warning system for air quality control and management, it is important to provide precise information about the future concentrations in pollutants. We present here a time series model in predicting the Air Pollution Index (API) from three different stations; industrial, residential, and sub-urban areas between 2000 and 2009. In this paper, the Box–Jenkins approach of seasonal autoregressive integrated moving average (ARIMA), artificial neural network (ANN), and three models of fuzzy time series (FTS) have been compared by using the mean absolute percentage error, mean absolute error, mean square error, and root mean square error. Although all the methods were used as operational tools, the ANN seemed more accurate in forecasting API. The results showed that FTS (i.e. Chen’s, Yu’s, and Cheng’s) performed inconsistent results since the conventional methods of ARIMA outperformed the performance of FTS. However, consistent results were achieved as the ANNs gave the smallest forecasting error compared to FTS and ARIMA.

Aiming at a large number of ambiguous, imprecise and incomplete data in the real world, fuzzy time series has come into being and developed into an effective forecasting approach. In the process of modeling and forecasting of fuzzy time series, the prediction performance of fuzzy time series can be effectively improved by partitioning the universe of discourse into different lengths. In this paper, a forecasting approach for fuzzy time series, which introduces the granularity mechanism into interval division and employs differential data for incremental forecasting, is proposed to solve the problem of time series forecasting with high forecasting precision. In the proposed approach, in order to describe the fuzzy logic relationship and fuzzy trend of historical data, we first do differential processing on the historical samples. Then, Fuzzy C-means (FCM) clustering algorithm is used to generate several partition intervals tentatively. In the sequel, we use the principle of justifiable granularity to constantly adjust the width of all the intervals, so that these information granules associated with corresponding intervals become the most "informative" information granules. Finally, the boundary of information granules is used as the basis of interval division to complete the forecasting task. An illustrative example is provided to demonstrate the essence of the proposed approach. The comparative experiment with other representative approaches shows that the proposed approach can significantly improve the prediction accuracy of time series.

The application of fuzzy time series is used to view the stock of clothing materials. As for the problem so far, CV Duta Express does not have a model for seeing the stock of complete materials in the warehouse, so the process is not optimal. This will have an impact on orders that come in at the same time and in large quantities. to avoid stock shortages, which resulted in the company experiencing losses. The purpose of this study is to make it easier to predict the stock of clothing materials and to be able to analyze every stock management at CV Duta Express with a fuzzy time series model. The variables used are stock needs, the amount of stock of school clothes, batik clothes, and pants. The research methodology in data collection consists of product type data from 2018–2021 at CV Duta Express Aceh Utara. Data analysis needs consist of school clothes, school pants, and clothes. Next, the fuzzy time series process determines the actual data is the type of school clothes, and what is forecast is sales for January 2018–2021 at the end of December. For a value of 1108 A2 fuzzification value, 172 A4 fuzzification value for batik clothes, and 894 pants with an A1 fuzzification value, then the value of the universe set used is U = [26, 323]. The value of forming a linguistic set is based on the length of the interval U3 = [111,153], U4 = [153,196], and U5 = [196,238]. The result of the fuzzification value from historical data for the value of 172 fuzzification is A4, for data of 133 fuzzification is A3. The formation of Fuzzy Logic Relationship (FLR) values for the period 1/7/2021 to 1/13/2021 is obtained from the data range A4=174 and range A3=125 in each period to be related. The results of forecasting with fuzzy time series testing at the end of December 2021 are 196 stocks of clothes that must be optimized in the following month. The test results in this study are to see if the error value using the AFER model is 0.4511% while the RMSE test value has an error value of 5.0199. After being calculated for the forecast every month, the average obtained for AFER is 0.50154 % and the RMSE is 9.86518.. Keywords : Forecasting, Fuzzy Time Series, stock

Demak Regency is one of the regions in Central Java Province with a low incidence of Dengue Fever compared to other cities and districts. Even so, DHF control needs to be done to minimize the occurrence of dengue fever, because DHF is a fairly dangerous disease. One form of controlling the number of DHF events that is widely used is using forecasting models, one of them is using Fuzzy Time Series. The Multivariate Fuzzy Time Series (MFTS) model is a development of the Fuzzy Time Series model that can be used to forecast using time series data by using more than one variable for forecasting, compared to the Fuzzy Time Series method that usually using only one variable. Based on the research results obtained, the MFTS model has a fairly accurate MAPE value, wherein the best MAPE was at 3 years scenario with MAPE 10,728%.

Web-based predictions have an important role in economic, health and the environment issues. Prediction by using Fuzzy Time Series model has advantage in the use of historical data that classifies the set data into a blur limitation (fuzzy) compared to other conventional time series model which use data firm (crisp). As weather prediction that can give impact to humans, the prediction of pollutants PM10 concentration in the air also useful because the concentration of pollutant PM10 in the air is able to drive bad impact to humans. This research discusses about web-based prediction of PM10 pollutant concentration using Ruey Chyn Tsaur Fuzzy Time Series models. The pollutant PM10 concentrations data results obtained from the sensor will be sent to the web server and stored in an online database for further prediction using Fuzzy Time Series Ruey Chyn Tsaur models which implement re-divide the partition of interval after first partition of the universe of discourse. By using 30 data samples of pollutants PM10 concentrations in the prediction, we get the value of Mean Absolute Percentage Error (MAPE) is 0.0187%. It indicates that the Fuzzy Time Series Ruey Chyn Tsaur models is well proven to be used in predicting the pollutants PM10 concentration.

Stock market forecasting by using a hybrid model of exponential fuzzy time series

This paper introduces an integrated algorithm for forecasting electricity consumption (EL) based on fuzzy regression, time series and principal component analysis (PCA) in uncertain markets such as Iran. The algorithm is examined by mean absolute percentage error, analysis of variance (ANOVA) and Duncan Multiple Range Test. PCA is used to identify the input variables for the fuzzy regression and time series models. Monthly EL in Iran is used to show the superiority of the algorithm. Moreover, it is shown that the selected fuzzy regression model has better estimated values for total EL than time series. The algorithm provides as good results as intelligent methods. However, it is shown that the algorithm does not require utilization of preprocessing methods but genetic algorithm, artificial neural network and fuzzy inference system require preprocessing which could be a cumbersome task to deal with ambiguous data. The unique features of the proposed algorithm are three fold. First, two type of fuzzy regressions with and without preprocessed data are prescribed by the algorithm in order to minimize the bias. Second, it uses PCA approach instead of trial and error method for selecting the most important input variables. Third, ANOVA is used to statistically compare fuzzy regression and time series with actual data.

This study aims to (i) develop predictive models using statistical and fuzzy approaches, and (ii) evaluate their forecasting performance. The data were obtained from www.investing.com for the period 1961 to 2017 and focus on five ASEAN countries: Indonesia, Malaysia, the Philippines, Singapore, and Thailand. The statistical models used are Autoregressive Integrated Moving Average (ARIMA) and Exponential Smoothing, while the fuzzy approaches include Chen and Singh fuzzy time series models. The dataset was divided into training and test sets in a 75%-25% proportion. ARIMA models capture trends and autocorrelations in time series data, while Exponential Smoothing uses exponentially weighted averages. Fuzzy models are designed to handle uncertainty and linguistic patterns in data. The results show that Singh’s fuzzy model yields the lowest error for Indonesia, while exponential smoothing and Chen fuzzy time series model demonstrate the same lowest error for Malaysia. For the Philippines, exponential smoothing is most accurate, whereas ARIMA and Singh fuzzy time series achieve the smallest error for Singapore. For Thailand, exponential smoothing and ARIMA perform equally well. However, the robustness of the forecasting model cannot be determined from either statistical or fuzzy methods, highlighting the challenge in determining the most robust model for inflation in the ASEAN region. The 75%-25% data split may also limit the generalizability of the findings. This study contributes a rare cross-country comparison of statistical and fuzzy forecasting methods in the ASEAN context. It highlights the importance of model selection based on country-specific inflation behavior and provides insights for improving forecasting strategies in macroeconomic applications.

This paper is aimed at improving the forecasting accuracy with correcting two deficiencies, subintervals failing to well represent the data distribution structures and a single antecedent factor in the fuzzy relationships in current fuzzy time series models. First, the universe of discourse is partitioned into subintervals with the midpoints of two adjacent cluster centers generated by the fuzzy clustering method as their endpoints. And the sub-intervals are employed to fuzzify the time series into fuzzy time series. Then, the fuzzy time series model with multi-factors high-order fuzzy relationships is built up to forecast the stock markets. Finally, the model we produced is used to forecast the daily Shanghai Stock Exchange Composite index and Shenzhen Stock Exchange Component index, respectively. The results show that the model do improve the prediction accuracy compared with the benchmark model.

A drawback of traditional forecasting methods is that they can not deal with forecasting problems in which the historical data are represented by linguistic values. Using fuzzy time series to deal with forecasting problems can overcome this drawback. In this paper, we propose a new fuzzy time series model called the two-factors time-variant fuzzy time series model to deal with forecasting problems. Based on the proposed model, we develop two algorithms for temperature prediction. Both algorithms have the advantage of obtaining good forecasting results.

Natural rubber is one of the pillars of Indonesia's export commodities. However, over the last few years, the export value of natural rubber has decreased due to an oversupply of this commodity in the global market. To overcome this problem, it is possible to predict the volume of Indonesia natural rubber exports. Predicted values can also help the government to compile market intelligence for natural rubber commodities periodically. In this study, the prediction of the export volume of natural rubber was carried out using the Automatic Clustering as an interval maker in the Fuzzy Time Series or usually called Automatic Clustering and Fuzzy Logical Relationship (ACFLR). The data used is 51 data per year from 1970 to 2020. The purpose of this study is to predict the volume of Indonesia natural rubber exports and compare the prediction results between the Automatic Clustering and Fuzzy Logical Relationship (ACFLR) and Chen's Fuzzy Time Series. The results showed that there was a significant difference between the two methods, ACFLR got 0.5316% MAPE with and Chen's Fuzzy Time Series model got 8.009%. Show that the ACFLR method performs better than the pure Fuzzy Time Series in predicting volume of Indonesia natural rubber exports.