English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

This study aim to compare Artificial Neural Networks (ANNs), namely the Multilayer Perceptron, the Focused Time Lagged Feedforward Network, the Radial Basis Function network and Layer-Recurrent Neural network for forecasting financial time series, notably the future value of shares of the capital market. It was made use of MATLAB software for simulations. An experiment was performed aiming to evaluate the networks in various situations, especially with regard to the number of neurons in the hidden layer and the variation of the initial weights. For the simulations were chosen ten shares of the largest weight in the Bovespa index: Vale (VALE5), Petrobras (PETR4), Itaú-Unibanco (ITUB4), Bradesco (BBDC4), Vale (VALE3), AMBEV (ABEV3), Petrobras (PETR3), Itausa (ITSA4), Banco do Brasil (BBAS3) e BM&FBOVESPA (BVMF3). The input of the network were the historical prices of the chosen actions. The analysis was performed by comparing the values predicted by the ANNs with the real historical values, measured by the mean square error. These results are presented very satisfactory and the minimum errors, of the order of 10, was given by the Layer-Recurrent neural network. Key-words: Artificial Neural Networks. Time Series Forecasting. Stock Market.

ESTUDO COMPARATIVO DE REDES NEURAIS ARTIFICIAIS PARA PREVISÃO DE SÉRIES TEMPORAIS FINANCEIRAS