SAHAANN: A NOVEL EVOLUTIONARY ARTIFICIAL NEURAL NETWORK FOR IMPROVED FINANCIAL TIME SERIES FORECASTING
Here, we introduce SAHAANN, a new kind of hybrid forecasting model that combines ANN with SAHA, a recently created meta-heuristic, and other artificial neural network (ANN) techniques. Optimizing the ANN's weights and biases is what SAHA does. We tested this concept by simulating two popular st...
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
University of Kragujevac
2025-03-01
|
| Series: | Proceedings on Engineering Sciences |
| Subjects: | |
| Online Access: | https://pesjournal.net/journal/v7-n1/66.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Here, we introduce SAHAANN, a new kind of hybrid forecasting model that combines ANN with SAHA, a recently created meta-heuristic, and other artificial neural network (ANN) techniques. Optimizing the ANN's weights and biases is what SAHA does. We tested this concept by simulating two popular stock indices. To measure how well the model works, we mostly utilize MSE, RMSE, and MAPE. To pass the test, all you have to do is guess how much each stock index will be worth in one day and 10 days. We were able to see how the results of training the ANN model with different metaheuristics, such as the genetic algorithm (GA), particle swarm optimization (PSO), differential evolution (DE), fireworks algorithm (FWA), and chemical reaction optimization (CRO). In order, they are GAANN, PSOANN, DEANN, FWAANN, and CROANN. For every model, we perform an exhaustive evaluation. The SAHAANN model has a perfect record of success in the lab. |
|---|---|
| ISSN: | 2620-2832 2683-4111 |