Forecasting economic time series using chaotic neural networks

Abstract

This paper describes the application of KIII, a biologically more plausible neural network model, for forecasting economic time series. K-sets are connectionist models based on neural populations and have been used in many machine learning applications. In this paper, this method was applied to IPCA, a Brazilian consumer price index surveyed by IBGE. The values ranged from August 1994 to June 2017. Experiments were performed using four non-parametric models and seven parametric methods. The statistical metric RMSE was used to compare methods performance. Freeman KIII sets worked well as a filter, but it was not a good prediction method. This paper contributes with the use of non-parametrics models for forecasting inflation in a developing country.

Previsão de receitas de ICMS do estado do Espírito Santo através de Seleção de Características em Cascata e técnicas de Aprendizado de Máquina

“…Problemas de predição em séries temporais são desafiadores e tem sido estudados pela comunidade acadêmica. Revisitar técnicas de predição e adaptá-las a problemas concretos é sempre uma abordagem de interesse de pesquisa [Castanho 2011, de Azevedo et al 2017, Makridakis et al 2018b, Gonçalves and Rosa 2018, Domingos Jr et al 2018. Além disso, existem as surpresas que, apesar de serem fortemente estudadas pelo mercado e por pesquisadores [Caruso 2019], são dificilmente previsíveis pela sua natureza.…”