MOBILE ALTERNATIVE OF THE MONEYBEE PROJECT FOR FINANCIAL FORECASTING
MOBILE ALTERNATIVE OF THE MONEYBEE PROJECT FOR FINANCIAL FORECASTING

Author(s): Petar Tomov, Iliyan Zankinski, Maria Barova
Subject(s): Geography, Regional studies
Published by: Сдружение „Академия за иновации и устойчивост“
Keywords: artificial neural networks; mobile computing; time series forecasting
Summary/Abstract: In an attempt to reduce the time for processing of huge volumes of data and algorithms calculations researchers have been addressing the concept of distributed computing for the last couple of decades [4,5]. The possibility of utilising the process power of systems in their idle mode has raised yet another much valuable and largely applicable concept, namely donated distributed computing. According to DistributedComputing.Info website, there are hundreds of distributed computing projects in many scientific areas. The most interesting fact, clearly visible in the website, is that there is only one active financial project called MQL5 Cloud Network. Between 2000 and 2010 a German project called MoneyBee was under study. The aim of this project was financial forecasting by artificial neural networks which were trained in a donated distributed computing environment. MoneyBee project was developed and operated by i42 Informations management GmbH company. It was a distributed solution based on a central server and many remote workers. Calculations performed on workers were done during the desktop idle mode while a screensaver was running. In fact, MoneyBee on client machines was the screensaver itself. The goal of the current research is to recreate the ideas used in MoneyBee project on modern mobile devices. Android Live Wallpaper is employed for background calculations executed for training of artificial neural network [9] in the mobile systems idle mode. For the needs of the study two GitHub projects are used - VitoshaTrade and VitDisComp.

• Details
• Contents