Home > Research > Publications & Outputs > A new intuitionistic fuzzy functions approach b...

Links

Text available via DOI:

View graph of relations

A new intuitionistic fuzzy functions approach based on hesitation margin for time-series prediction

Research output: Contribution to journalJournal articlepeer-review

E-pub ahead of print

Standard

A new intuitionistic fuzzy functions approach based on hesitation margin for time-series prediction. / Cagcag Yolcu, Ozge; Bas, Eren; Egrioglu, Erol; Yolcu, Ufuk.

In: Soft Computing, 01.11.2019.

Research output: Contribution to journalJournal articlepeer-review

Harvard

APA

Vancouver

Author

Bibtex

@article{bff6fc63ffbd47c6a66dd04f8b0d045e,
title = "A new intuitionistic fuzzy functions approach based on hesitation margin for time-series prediction",
abstract = "There are various studies in which a variety of prediction tools have been introduced in time-series prediction literature. Non-probabilistic approaches which are based on fuzzy set theory, especially in recent years, have been put forward. Although these approaches including adaptive network fuzzy inference system, fuzzy functions approach, and fuzzy regression can be successfully utilized as a prediction tool, they have not been designed for prediction problem and they pass over the dependency structure of time-series observations. From this point forth, designing a prediction tool that considers the dependency structure of the observations of time series will procure to get predictions more accurately. Although the membership values, in the analysis process, are taken into account in almost all fuzzy methods, the non-membership and hesitation values are not considered. However, using as much information as possible on time series may be another positive factor that gives more accurate predictions. The primary aim of this study, for time-series prediction, is to introduce an intuitionistic fuzzy regression functions approach based on hesitation margin (I-FRF-HM). In the introduced intuitionistic fuzzy regression functions approach, two inference systems are separately constituted such that while one of them uses membership, other one uses non-membership values as inputs of inference system in addition with the crisp observations of time series. Predictions obtained from each system are converted into final predictions of whole inference system via an approach based on hesitation margin. Intuitionistic fuzzy C-means are utilized to get membership and non-membership values in the proposed model. The proposed I-FRF-HM has been applied to various real-world time series. The obtained findings are evaluated along with the results of some other time-series prediction models. The results show that the proposed I-FRF-HM has superior prediction performance to other prediction models.",
keywords = "Fuzzy regression functions, Hesitation margin, Intuitionistic fuzzy C-means, Time-series prediction",
author = "{Cagcag Yolcu}, Ozge and Eren Bas and Erol Egrioglu and Ufuk Yolcu",
year = "2019",
month = nov,
day = "1",
doi = "10.1007/s00500-019-04432-2",
language = "English",
journal = "Soft Computing",
issn = "1432-7643",
publisher = "Springer",

}

RIS

TY - JOUR

T1 - A new intuitionistic fuzzy functions approach based on hesitation margin for time-series prediction

AU - Cagcag Yolcu, Ozge

AU - Bas, Eren

AU - Egrioglu, Erol

AU - Yolcu, Ufuk

PY - 2019/11/1

Y1 - 2019/11/1

N2 - There are various studies in which a variety of prediction tools have been introduced in time-series prediction literature. Non-probabilistic approaches which are based on fuzzy set theory, especially in recent years, have been put forward. Although these approaches including adaptive network fuzzy inference system, fuzzy functions approach, and fuzzy regression can be successfully utilized as a prediction tool, they have not been designed for prediction problem and they pass over the dependency structure of time-series observations. From this point forth, designing a prediction tool that considers the dependency structure of the observations of time series will procure to get predictions more accurately. Although the membership values, in the analysis process, are taken into account in almost all fuzzy methods, the non-membership and hesitation values are not considered. However, using as much information as possible on time series may be another positive factor that gives more accurate predictions. The primary aim of this study, for time-series prediction, is to introduce an intuitionistic fuzzy regression functions approach based on hesitation margin (I-FRF-HM). In the introduced intuitionistic fuzzy regression functions approach, two inference systems are separately constituted such that while one of them uses membership, other one uses non-membership values as inputs of inference system in addition with the crisp observations of time series. Predictions obtained from each system are converted into final predictions of whole inference system via an approach based on hesitation margin. Intuitionistic fuzzy C-means are utilized to get membership and non-membership values in the proposed model. The proposed I-FRF-HM has been applied to various real-world time series. The obtained findings are evaluated along with the results of some other time-series prediction models. The results show that the proposed I-FRF-HM has superior prediction performance to other prediction models.

AB - There are various studies in which a variety of prediction tools have been introduced in time-series prediction literature. Non-probabilistic approaches which are based on fuzzy set theory, especially in recent years, have been put forward. Although these approaches including adaptive network fuzzy inference system, fuzzy functions approach, and fuzzy regression can be successfully utilized as a prediction tool, they have not been designed for prediction problem and they pass over the dependency structure of time-series observations. From this point forth, designing a prediction tool that considers the dependency structure of the observations of time series will procure to get predictions more accurately. Although the membership values, in the analysis process, are taken into account in almost all fuzzy methods, the non-membership and hesitation values are not considered. However, using as much information as possible on time series may be another positive factor that gives more accurate predictions. The primary aim of this study, for time-series prediction, is to introduce an intuitionistic fuzzy regression functions approach based on hesitation margin (I-FRF-HM). In the introduced intuitionistic fuzzy regression functions approach, two inference systems are separately constituted such that while one of them uses membership, other one uses non-membership values as inputs of inference system in addition with the crisp observations of time series. Predictions obtained from each system are converted into final predictions of whole inference system via an approach based on hesitation margin. Intuitionistic fuzzy C-means are utilized to get membership and non-membership values in the proposed model. The proposed I-FRF-HM has been applied to various real-world time series. The obtained findings are evaluated along with the results of some other time-series prediction models. The results show that the proposed I-FRF-HM has superior prediction performance to other prediction models.

KW - Fuzzy regression functions

KW - Hesitation margin

KW - Intuitionistic fuzzy C-means

KW - Time-series prediction

U2 - 10.1007/s00500-019-04432-2

DO - 10.1007/s00500-019-04432-2

M3 - Journal article

AN - SCOPUS:85074694938

JO - Soft Computing

JF - Soft Computing

SN - 1432-7643

ER -