Forecasting and forecast-combining of quarterly earnings-per-share via genetic programming

In this study we examine different methodologies to estimate earnings. More specifically, we evaluate the viability of Genetic Programming as both a forecasting model estimator and a forecast-combining methodology. When we compare the performance of traditional mechanical forecasting (ARIMA) models and models developed using Genetic Programming we observe that Genetic Programming can be used to create time-series models for quarterly earnings as accurate as the traditional linear models. Genetic Programming can also effectively combine forecasts. However, Genetic Programming's forecast combinations are sometimes unable to improve on Value Line. Moreover, simple averaging of forecasts results in better predictive accuracy than Genetic Programming-combining of forecasts. Hence, as implemented in this study, Genetic Programming is not superior to traditional methodologies in either forecasting or forecast combining of quarterly earnings.

Abstract licence: CC BY-NC-SA

12 more references, whose wording the paper’s licence keeps private.

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.