Search results

A significant number of studies have been conducted to analyze and understand the relationship between gas emissions and global temperature using conventional statistical approaches. However, these techniques follow assumptions of probabilistic modeling, where results can be associated with large errors. Furthermore, such traditional techniques cannot be applied to imprecise data. The purpose of this paper is to avoid strict assumptions when studying the complex relationships between variables by using the three innovative, up-to-date, statistical modeling tools: adaptive neuro-fuzzy inference systems (ANFIS), artificial neural networks (ANNs) and fuzzy time series models.

These three approaches enabled us to effectively represent the relationship between global carbon dioxide (CO₂) emissions from the energy sector (oil, gas and coal) and the average global temperature increase. Temperature was used in this study (1900-2012). Investigations were conducted into the predictive power and performance of different fuzzy techniques against conventional methods and among the fuzzy techniques themselves.

A performance comparison of the ANFIS model against conventional techniques showed that the root means square error (RMSE) of ANFIS and conventional techniques were found to be 0.1157 and 0.1915, respectively. On the other hand, the correlation coefficients of ANN and the conventional technique were computed to be 0.93 and 0.69, respectively. Furthermore, the fuzzy-based time series analysis of CO₂ emissions and average global temperature using three fuzzy time series modeling techniques (Singh, Abbasov–Mamedova and NFTS) showed that the RMSE of fuzzy and conventional time series models were 110.51 and 1237.10, respectively.

The paper provides more awareness about fuzzy techniques application in CO₂ emissions studies.

These techniques can be extended to other models to assess the impact of CO₂ emission from other sectors.

Many studies have proposed variant fuzzy time series models for uncertain and vague data. The purpose of this paper is to adapt a fuzzy time series combined with genetic algorithm (GA) to forecast tourist arrivals in Taiwan.

Different cases are studied to understand the effect of variation of fuzzy time series order, number of intervals and population size on the fitness function which decreases with increase in fuzzy time series order and number of fuzzy intervals, but do not have marginal effect due to change in population size.

Results based on an example of forecasting Taiwan’s tourism demand was used to verify the efficacy of proposed model and confirmed its superiority to existing models providing solutions for different orders of fuzzy time series, number of intervals and population size with a smaller forecasting error as measured by root mean square error.

This study provides a viable forecasting methodology, adapting a fuzzy time series combined with an evolutionary GA. The proposed hybridized framework of fuzzy time series and GA, where GA is used to calibrate fuzzy interval length, is flexible and replicable to many industrial situations.

In the last few decades, there has been growing interest in forecasting with computer intelligence, and both fuzzy time series (FTS) and artificial neural networks (ANNs) have gained particular popularity, among others. Rather than the conventional methods (e.g., econometrics), FTS and ANN are usually thought to be immune to fundamental concepts such as stationarity, theoretical causality, post-sample control, among others. On the other hand, a number of studies significantly indicated that these fundamental controls are required in terms of the theory of forecasting, and even application of such essential procedures substantially improves the forecasting accuracy. The aim of this paper is to fill the existing gap on modeling and forecasting in the FTS and ANN methods and figure out the fundamental concepts in a comprehensive work through merits and common failures in the literature. In addition to these merits, this paper may also be a guideline for eliminating unethical empirical settings in the forecasting studies.

This study aims to analyze the performance of four different fuzzy inference system-based forecasting tools using a real case company.

The forecasting tools were tested using 27 products of the nail polish line of a multinational beauty company and the performance of said tools was compared to those of the company’s previous forecasting methods that were basically qualitative (informal and intuition-based).

The performance of the methods analyzed was compared by using mean absolute percentage error. It was possible to determine the characteristics and conditions that make each model the best for each situation. The main takeaways were that low kurtosis, negatively skewed demand time-series and longer horizon forecasts that favor the fuzzy inference system-based models. Besides, the results suggest that the fuzzy forecasting tools should be preferred for longer horizon forecasts over informal qualitative methods.

Notwithstanding the proposed hybrid modeling approach based on fuzzy inference systems, our research offers a relevant contribution to theory and practice by shedding light on the segmentation and selection of forecasting models, both in terms of time-series characteristics and forecasting horizon. The proposed fuzzy inference systems showed to be particularly useful not only when time-series distributions present no clear central tendency (that is, they are platykurtic or dispersed around a large plateau around the median, which is the characteristic of negative kurtosis), but also when mode values are greater than median values, which in turn are greater than mean values. This large tail to the left (negative skewness) is typical of successful products whose sales are ramping up in early stages of their life cycle. For these, fuzzy inference systems may help managers screen out forecast bias and, therefore, lower forecast errors. This behavior also occurs when managers deal with forecasts of longer horizons. The results suggest that further research on fuzzy inference systems hybrid approaches for forecasting should emphasize short-term forecasting by trying to better capture the “tribal” managerial knowledge instead of focusing on less dispersed and slower moving products, where the purely qualitative forecasting methods used by managers tend to perform better in terms of their accuracy.

The purpose of this paper is to analyze demand forecast strategies to support a more sustainable management in a pallet supply chain, and thus avoid environmental impacts, such as reducing the consumption of forest resources.

Since the producer presents several uncertainties regarding its demand logs, a methodology that embed zero-inflated intelligence is proposed combining fuzzy time series with clustering techniques, in order to deal with an excessive count of zeros.

A comparison with other models from literature is performed. As a result, the strategy that considered at the same time the excess of zeros and low demands provided the best performance, and thus it can be considered a promising approach, particularly for sustainable supply chains where resources consumption is significant and exist a huge variation in demand over time.

The findings of the study contribute to the knowledge of the managers and policymakers in achieving sustainable supply chain management. The results provide the important concepts regarding the sustainability of supply chain using fuzzy time series and clustering techniques.

The purpose of this paper is to develop a novel intuitionistic fuzzy seasonality regression (IFSR) with particle swarm optimization (PSO) algorithms to accurately forecast air pollutions, which are typical seasonal time series data. Seasonal time series prediction is a critical topic, and some time series data contain uncertain or unpredictable factors. To handle such seasonal factors and uncertain forecasting seasonal time series data, the proposed IFSR with the PSO method effectively extends the intuitionistic fuzzy linear regression (IFLR).

The prediction model sets up IFLR with spreads unrestricted so as to correctly approach the trend of seasonal time series data when the decomposition method is used. PSO algorithms were simultaneously employed to select the parameters of the IFSR model. In this study, IFSR with the PSO method was first compared with fuzzy seasonality regression, providing evidence that the concept of the intuitionistic fuzzy set can improve performance in forecasting the daily concentration of carbon monoxide (CO). Furthermore, the risk management system also implemented is based on the forecasting results for decision-maker.

Seasonal autoregressive integrated moving average and deep belief network were then employed as comparative models for forecasting the daily concentration of CO. The empirical results of the proposed IFSR with PSO model revealed improved performance regarding forecasting accuracy, compared with the other methods.

This study presents IFSR with PSO to accurately forecast air pollutions. The proposed IFSR with PSO model can efficiently provide credible values of prediction for seasonal time series data in uncertain environments.

There is a growing interest in fuzzy time series (FTS) forecasting, and several improvements are presented in the last few decades. Among these improvements, the development of causal models (i.e., multiple factor FTS) has sparked a particular literature dealing with the causal inference and its integration in the FTS framework. However, causality among variables is usually introduced as a subjective assumption rather than empirical evidence. As a result of arbitrary causal modeling, the existing multiple factor FTS models are developed with implicit forecasting failure. Since post-sample control (unknown future, as in the business practice) is usually ignored, the spurious accuracy gain through increasing factors is not identified by scholars. This paper discloses the use of causality in the FTS method, and investigates the spurious causal inference problem in the literature with a justification approach. It invalidates the contribution of dozens of previously published papers while justifying its claim with illustrative examples and a comprehensive set of experiments with random data, as well as real business data from maritime transportation (Baltic Dry Index).

This study aims to adapt a neural network based fuzzy time series model to improve Taiwan's tourism demand forecasting.

Fuzzy sets are for modeling imprecise data and neural networks are for establishing non‐linear relationships among fuzzy sets. A neural network based fuzzy time series model is adapted as the forecasting model. Both in‐sample estimation and out‐of‐sample forecasting are performed.

This study outperforms previous studies undertaken during the SARS events of 2002‐2003.

The forecasting model only takes the observation of one previous time period into consideration. Subsequent studies can extend the model to consider previous time periods by establishing fuzzy relationships.

Non‐linear data is complicated to forecast, and it is even more difficult to forecast nonlinear data with shocks. The forecasting model in this study outperforms other studies in forecasting the nonlinear tourism demands during the SARS event of November 2002 to June 2003.

The purpose of this paper is to test the effectiveness of fuzzy time series (FTS) forecasting system in a supply chain experiencing disruptions and also to examine the changes in performance as the authors move across different tiers.

A discrete event simulation based on the popular beer game model is used for these tests. A popular ordering management system is used to emulate the behavior of the system when the game is played with human players.

FTS is tested against some other well-known forecasting systems and it proves to be the best of the lot. It is also shown that it is better to go for higher order FTS for higher tiers, to match auto regressive integrated moving average.

This study fills an important research gap by proving that FTS forecasting system is the best for a supply chain during disruption scenarios. This is important because the forecasting performance deteriorates significantly and the effect is more pronounced in the upstream tiers because of bullwhip effect.

Having a system which works best in all scenarios and also across the tiers in a chain simplifies things for the practitioners. The costs related to acquiring and training comes down significantly.

This study contributes by suggesting a forecasting system which works best for all the tiers and also for every scenario tested and simultaneously significantly improves on the previous studies available in this area.

The purpose of this paper is to provide a better method for quality management to maintain an essential level of quality in different fields like product quality, service quality, air quality, etc.

In this paper, a hybrid adaptive time-variant fuzzy time series (FTS) model with genetic algorithm (GA) has been applied to predict the air pollution index. Fuzzification of data is optimized by GAs. Heuristic value selection algorithm is used for selecting the window size. Two algorithms are proposed for forecasting. First algorithm is used in training phase to compute forecasted values according to the heuristic value selection algorithm. Thus, obtained sequence of heuristics is used for second algorithm in which forecasted values are selected with the help of defined rules.

The proposed model is able to predict AQI more accurately when an appropriate heuristic value is chosen for the FTS model. It is tested and evaluated on real time air pollution data of two popular tourism cities of India. In the experimental results, it is observed that the proposed model performs better than the existing models.

The management and prediction of air quality have become essential in our day-to-day life because air quality affects not only the health of human beings but also the health of monuments. This research predicts the air quality index (AQI) of a place.

The proposed method is an improved version of the adaptive time-variant FTS model. Further, a nature-inspired algorithm has been integrated for the selection and optimization of fuzzy intervals.