Forecasting EPS o with Hybrid Genetic algorithm, particle swarm optimization and Neural networks
Forecasting earnings per share (EPS) are among the most important and crucial tasks for both outside investors and internal managers. The focus of most articles in literature is forecasting EPS with linear methods. Researchers rarely employ nonlinear models to forecast EPS. However some researchers...
Guardado en:
| Autores principales: | , |
|---|---|
| Formato: | article |
| Lenguaje: | FA |
| Publicado: |
Shahid Bahonar University of Kerman
2017
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/9bae08e4e525460787f5b4a6d451c73a |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | Forecasting earnings per share (EPS) are among the most important and crucial tasks for both outside investors and internal managers. The focus of most articles in literature is forecasting EPS with linear methods. Researchers rarely employ nonlinear models to forecast EPS. However some researchers show that nonlinearities exist in the relation between EPS and its determinants. This finding of nonlinearities provides support for the use of non-linear models in the field of earnings per share.In this paper, the model based on an Artificial Neural Network (ANN) to predict EPS is proposed. After that, ANN model was optimized by Genetic Algorithm and particle swarm optimization. The Genetic Algorithm and particle swarm optimizations used to select the most relevant input variables because selection of input variables is a key stage in building predictive models, In this paper, The Genetic Algorithm and particle swarm optimizations used to select the effective variables on the EPS using a sample of 131 companies listed in the Tehran Stock Exchange through for the years 1389_1391.The results reveals that the genetic algorithm and swarm optimization is able to extract the effective variables on the EPS amongst the factors affecting EPS, and also have been caused improve power capabilities and expansion of neural network structure. |
|---|