A Comparative Study Of Backpropagation Algorithms In Financial Prediction

  title={A Comparative Study Of Backpropagation Algorithms In Financial Prediction},
  author={Salim Lahmiri},
  journal={International Journal of Computer Science, Engineering and Applications},
  • S. Lahmiri
  • Published 30 August 2011
  • Business
  • International Journal of Computer Science, Engineering and Applications
Stock market price index prediction is a challenging task for investors and scholars. Artificial neural networks have been widely employed to predict financial stock market levels thanks to their ability to model nonlinear functions. The accuracy of backpropagation neural networks trained with different heuristic and numerical algorithms is measured for comparison purpose. It is found that numerical algorithm outperform heuristic techniques. 

Figures and Tables from this paper

Improving the genetic-algorithm-optimized wavelet neural network for stock market prediction
This paper improves stock market prediction based on genetic algorithms (GA) and wavelet neural networks (WNN) and reports significantly better accuracies compared to existing approaches to stock
Evolving Stock Market Prediction Models Using Soft Computing Techniques
A new market forecasting model based on soft computing and especially genetic progra mming is developed to enhance the investor decision and suggests that the stock market can be forecast using soft computing technique.
A comparative study of predictive algorithms for time series forecasting
A comparative study of three predictive algorithms: Linear Regression, Support Vector Machines and Multilayer Perceptron in order to determine their performances in term of implementing times series for predictive systems exhibits that linear regression produced the best forecasts.
Efficacy of a Classical and a Few Modified Machine Learning Algorithms in Forecasting Financial Time Series
Financial markets and economy forecast are closely related to each other. Forecast of prices of financial assets is therefore of importance for any economy-planning be it global, national or
ANN based method for improving gold price forecasting accuracy through modified gradient descent methods
Improvement in forecasting accuracy is achieved by proposing and developing a few modified GDM algorithms that incorporate different optimization functions by replacing the standard quadratic error function of classical GDM.
Bitcoin price prediction using ensembles of neural networks
  • Edwin Sin, Lipo Wang
  • Computer Science
    2017 13th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD)
  • 2017
The ensemble was used to predict the next day direction of the price of Bitcoin given a set of approximately 200 features of the cryptocurrency over a span of 2 years, and a trading strategy that follows the single, best MLP model in the ensemble generated approximately 53% in returns.
Comparison of Particle Swarm Optimization and Backpropagation Algorithms for Training Feedforward Neural Network
To improve the performance of ANN, the adjustment of network weights using Particle Swarm Optimization was proposed as a mechanism and the results showed that PSO approach outperforms the BP for training neural network models.
Forecasting demand for low cost carriers in Australia using an artificial neural network approach
AbstractThis study focuses on predicting Australia's low cost carrier passenger demand and revenue passenger kilometres performed (RPKs) using traditional econometric and artificial neural network
Electricity price forecasting in Turkey with artificial neural network models
The electrici efficiency has been found to be nonlinear pro neural netwo momentum, number of n compared ac observed as 9.
Short term load forecasting of Indian system using linear regression and artificial neural network
The hour ahead load forecasting is used for the reliable and proactive operation of the power system. The hour ahead load forecasting is a one type of Short Term Load Forecasting (STLF). The mostly


A Hybrid Neurogenetic Approach for Stock Forecasting
The neurogenetic hybrid showed notable improvement on the average over the buy-and-hold strategy and the context-based ensemble further improved the results, which implies that the proposed neurogenetics hybrid can be used for financial portfolio construction.
Trend time series modeling and forecasting with neural networks
  • M. Qi, G. Zhang
  • Computer Science
    2003 IEEE International Conference on Computational Intelligence for Financial Engineering, 2003. Proceedings.
  • 2003
This paper investigates how to best model trend time series using neural networks and finds that with neural networks differencing often gives meritorious results regardless of the underlying DGPs.
Conjugate gradient algorithm for efficient training of artifi-cial neural networks
A novel approach is presented for the training of multilayer feedforward neural networks, using a conjugate gradient algorithm, which is superior to that of the conventional backpropagation algorithm and is based on strong theoretical reasons supported by the numerical results of three examples.
Comparison of different back-propagation algorithms used in the diagnosis of transformer oil
  • L. Mokhnache, A. Boubakeur
  • Physics, Medicine
    Annual Report Conference on Electrical Insulation and Dielectric Phenomena
  • 2002
Back propagation Algorithm is applied in transformer oil diagnosis using Levenberg-Marquardt technique, Momentum technique and adaptive learning rate technique.
Performance comparison between backpropagation algorithms applied to intrusion detection in computer network systems
A topology of neural network intrusion detection system is proposed on which different backpropagation algorithms are benchmarked and the most optimum solution is selected with respect to mean square error.
Comparison of Neural Network Learning Algorithms for Prediction Enhancement of a Planning Tool
This work presents the results of the studies concerning the application of different neural network traini ng algorithms to enhance the prediction of a radio network planning tool. Investigations
Performance Comparison for MLP Networks Using Various Back Propagation Algorithms for Breast Cancer Diagnosis
The MLP network using the Levenberg-Marquardt algorithm displays the best performance for all the five measurement criteria's (accuracy, specificity, sensitivity, true positive and true negative) at a lower number of hidden nodes.
On the approximate realization of continuous mappings by neural networks
Learning representations by back-propagating errors
Back-propagation repeatedly adjusts the weights of the connections in the network so as to minimize a measure of the difference between the actual output vector of the net and the desired output vector, which helps to represent important features of the task domain.