Record Details

Study of some probabilistic and computational methods for fuzzy and intuitionistic fuzzy time series forecasting

KrishiKosh

View Archive Info
 
 
Field Value
 
Title Study of some probabilistic and computational methods for fuzzy and intuitionistic fuzzy time series forecasting
 
Creator Gangwar, Sukhdev Singh
 
Contributor Sanjay Kumar
 
Subject fuzzy logic, computer techniques, time series, forecasting
 
Description Thesis-PhD
Time series forecasting in the environment of fuzzy and intuitionistic finds its scope in various branches of sciences and engineering. The present research work mainly focus on development and implementation of fuzzy and intuitionistic fuzzy time series forecasting models. In this study, 6 different models of fuzzy and intuitionistic fuzzy time series forecasting were developed. Model 1, model 2 and model 3 were computational algorithm based higher order fuzzy time series forecasting models. These models were based on multiple partitions which were created using a ratio formula that uses maximum and minimum of time series data. Model 4, model 5 and model 6 used the concept of IFS to include degree of hesitation in time series forecasting. In model 4, the concept of IFS was used in fuzzification process. In model 5, fuzzy sets induced from IFS were used to establish fuzzy logical relations. Model 6 is purely intuitionistic fuzzy time series forecasting model. Both model 4 and model 5 used max-min composition of fuzzy logical relations and model 6 used max-min composition of intuitionistic fuzzy logical relations. All developed models (model 1 to model 6) were implemented on the time series data of enrollments of University of Alabama and share prices of SBI. Results of all forecasting models have been presented graphically to show their trend with actual time series data. Performance analysis of all models has been discussed in terms of MSE and AFE. Time series data of enrollments of University of Alabama were used to show robustness of all models to show their sensitivity towards the unexpected fluctuation in time series data. In forecasting enrollments of University of Alabama, the performance of model 3 in terms of both MSE and AFE is found better than model 1 and model 2. In forecasting share prices of SBI, the performance of model 1 in terms of both MSE and AFE is found better than model 2 and model 3. In forecasting enrollments of University of Alabama, the performance of model 5 in terms of both MSE and AFE is better than other IFS based forecasting models (model 4 and model 6). In forecasting share prices of SBI, the performance of model 6 in terms of MSE is found better than model 4 and model 5 but the performance of model 5 in terms of AFE is found better other model 4 and model 6. Overall, in the case of forecasting the enrollments of University of Alabama, model 3 outperforms than the model 1, model 2, model 4, model 5 and model 6 in terms of MSE and AFE. In terms of MSE and AFE, model 1 outperforms than the model 2, model 3, model 4, model 5 and model 6 in forecasting share prices of SBI. Model 6, intuitionistic fuzzy time series forecasting model and also gives the good performance with respect to other models. The present study may be helpful in research of many forecasting branches as well as in the fields of fuzzy and intuitionistic fuzzy time series forecasting.
 
Date 2016-11-04T13:55:59Z
2016-11-04T13:55:59Z
2013-08
 
Type Thesis
 
Identifier http://krishikosh.egranth.ac.in/handle/1/83921
 
Language en
 
Format application/pdf
 
Publisher G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand)