Neural network based hybrid prediction models for healthcare applications
Prediction models based on different concepts have been proposed in recent years. Improving the accuracy of prediction models has remained as a challenging task for researchers. In the development of time series analysis, it is well known that many phenomena are non-linear and hence only linear time...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Published: |
2012
|
| Subjects: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Prediction models based on different concepts have been proposed in recent years. Improving the accuracy of prediction models has remained as a challenging task for researchers. In the development of time series analysis, it is well known that many phenomena are non-linear and hence only linear time series prediction models are not sufficient to get accurate prediction. In the real-world, the time series data consist of complex linear and non-linear patterns and it may be difficult to obtain high prediction accuracy rates using only linear or only neural network model. Hybrid model which combines both linear and neural network models provides a better solution for time series prediction. In this study, new hybrid models termed as enhanced hybrid model (EHM), adaptive enhanced hybrid model (AEHM), dual enhanced hybrid model with fuzzy logic (DEHM-F), are proposed for univariate time series prediction. The proposed models take into account the pattern of data in selecting the best linear model and also in optimizing the configuration of the neural network. An enhanced adaptive neuro-fuzzy inference system (E-ANFIS) is also proposed for univariate time series prediction. E-ANFIS makes use of a strategy to determine the optimum number of input lags. |
|---|