Title of Paper:A Parallel Fuzzy Rule Discovery Algorithm and Future Goods Automated Trading System

Hits:

Date of Publication:2017-01-01

Included Journals:EI、CPCI-S、Scopus

Page Number:5652-5657

Key Words:Big data; Fuzzy rule based classification system; MapReduce; Future goods data

Abstract:Nowadays, with the rapid development of information technology, many business fields gradually usher in the era of big data. Extracting valuable information from the large data is very urgent. However, because of some limitations such as memory restrictions, data complexity or time complexity, the majority of traditional data mining methods are time consuming and low efficiency working on big data. In contrast, parallel computing is a common and reliable choice to solve that problem. In this paper, we propose a fuzzy rule discovery algorithm - fuzzy rule based classification system in the framework of MapReduce (FRBCS-MR). The proposed FRBCS-MR algorithm applies parallel computing technology to extract fuzzy rules and build fuzzy rule-base, which combines the advantage of dealing with uncertainty of fuzzy system and the ability of MapReduce in parallel computing. Future goods trading has been gradually transformed from the original manual trading to stylized trading and the stylized trading is considered to be the forefront and the most scientific mode of investment. In the experimental studies, the rules extracted by FRBCS-MR algorithm using future goods data are turned into trading strategies which are to be used in the real automated trading platform TradeBlazer (TB) to compute the profit and loss situation. The result shows that the FRBCS-MR algorithm has good profit ability and possesses a certain guiding significance for investors.