Development and performance evaluation of hybrid KELM models for forecasting of agro-commodity price
| Autor: | Nirjharinee Parida, Narendra Kumar Rout, Debahuti Mishra, Kaberi Das |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Artificial neural network
business.industry Computer science Cognitive Neuroscience Commodity Particle swarm optimization 020206 networking & telecommunications 02 engineering and technology Machine learning computer.software_genre Improved performance Mathematics (miscellaneous) Artificial Intelligence Kernel (statistics) Genetic algorithm 0202 electrical engineering electronic engineering information engineering Range (statistics) 020201 artificial intelligence & image processing Computer Vision and Pattern Recognition Artificial intelligence business computer Extreme learning machine |
| Zdroj: | Evolutionary Intelligence. 14:529-544 |
| ISSN: | 1864-5917 1864-5909 |
| Popis: | Incurring the benefits of kernel type-based extreme learning machine (KELM), this paper proposes a Grey wolf optimization based multiquadratic kernel KELM (GWO-KELM) models for effective forecasting of agro-commodity prices such as Chana and Barley. The objective is to obtain best possible weight between hidden and output layer of an artificial neural network using GWO optimization techniques. These weights are used in KELM model to achieve improved performance. To assess the superiority performance of the proposed models, similar results are obtained using simulation study by simulating GA-KELM (genetic algorithm based KELM), PSO-KELM (particle swarm optimization based KELM) and GWO-KELM multiquadratic kernel types (GWO-KELM) models. From exhaustive simulation result, it is demonstrated that GWO-KELM forecasting model offers the best performance using three models in terms of matrices. The study pertains to short and long range prediction for prices of Chana and Barley using publicly available database. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink