Deterministic fuzzy time series model for forecasting enrollments

Evolutionary Fuzzy Relational Modeling for Fuzzy Time Series Forecasting

“…To determine the learning parameters in the GA, we conduct a grid search of the population = {600, 800, 1000, 1200}, crossover rate = {0.6, 0.7, 0.8, 0.9}, and mutation rate = {0.01, 0.05, 0.10, 0.15}, and it reaches an optimal combination of population = 10,000, crossover rate = 0.9, and mutation rate = 0.1. The forecasting performance of the proposed method was also compared with those of other fuzzy time series approaches [3,4,10,29] using various metrics. The experimental results and the comparisons with other methods are shown in Table 6.…”

“…Li and Cheng [10] modified the matrix R based on a hidden Markov process. They took the repeated transitions into consideration when estimating the model parameters, and expressed the complete parameter set of the hidden Markov model.…”

“…Fuzzy relations are vital in constructing good fuzzy time series models, and such models can be categorized into two groups according to how they deal with these. The first group is implemented based on a relational matrix (e.g., [3,29], and our EFRM model), and the other is based on rule grouping (e.g., [4,10]). This study focuses on the first group, and thus how to construct the fuzzy relational matrix is the key issue with regard to obtaining better results in time series forecasting, and this is the focus of many related studies.…”

“…Although high-order models have been shown to improve forecasting accuracy, there is a significant problem with regard to determining an appropriate order number and order redundancy. However, this issue was recently resolved by Li and Cheng [10], who deterministically identified the fuzzy relationships without choosing an appropriate order number in advance. They proposed a deterministic one-factor high-order time-variant fuzzy time series model, which utilizes a backtrack algorithm to establish the fuzzy relationship rules, instead of repeating the same process used in earlier works.…”

A Hybrid Wind Speed Forecasting System Based on a ‘Decomposition and Ensemble’ Strategy and Fuzzy Time Series

“…Fuzzy time series are widely used in forecasting applications because of their capability of handling linguistic value datasets to obtain accurate forecasting. At present, it has been frequently and successfully used for forecasting nonlinear as well as dynamic datasets in various areas, including stock index [56], energy [57], course enrollment [58], green materials [59], load consumption [60], and so on. A fuzzy time series is defined by Song and Chissom [61] as follows.…”