Research On Fuzzy Rules Extraction Of Futures Trading Algorithm And Application Based On MapReduce

Fuzzy rule is an important basis for dealing with fuzzy problems and judging uncertain events.Fuzzy rule extracted approaches have been widely used in various field such as control,detection and finance fields.As the core algorithm of unsupervised learning in machine learning,clustering algorithm will play an important role in the extraction of fuzzy rules.For the purpose of deciding the trading signals in futures market,this paper proposes a clustering method to extract fuzzy rules of futures trading which can assist traders in trading.The futures market is booming,which makes the number of transactions increase gradually so that a huge amounts of data will be produced in the future.The proposed approach combines the calculation framework MapReduce to deal with futures data,which takes the advantage of parallel processing of experimental data without taking into account system details.Firstly,the MapReduce programming model is used to obtain the trend variable x(t)and the price difference ?p as the antecedent and consequent of the fuzzy rule by analyzing and dealing with the futures data.Secondly,the fuzzy sets of the antecedent and consequent of the rules are reasonably divided by combining the FCM algorithm.Thirdly,according to the sample membership degree distribution generates the fuzzy rule.Besides,meaningful fuzzy rules are extracted by reducing and selecting fuzzy rules based on calculating the confidence degree and support degree.Meanwhile,the profitability of the trading behavior can obtain from interpretable and understandable fuzzy rules which extracted by the proposed approach.Furthermore,comparing with the previous experience of the rules of futures trading,it can verify the validity and feasibility of the extracted If-Then fuzzy rules.The paper further improves the extraction method of fuzzy rules of futures trading by optimizing FCM algorithm,it not only explores the fuzzy rules but also improves the computational efficiency.And the paper will create a visualization page of fuzzy rules of futures trading by using the Angular front-end framework,which helps investors select satisfactory fuzzy rules of futures trading.Finally,the trading experiments are conducted on different futures products by applying the trading strategy which is modified based on the extracted fuzzy rules in Tradeblazer platform.And the improved strategy will be validated by the trading strategy performance test.The results show the strategy modified by the paper effectively increase the transaction success rate.So the fuzzy rules of futures trading extractedfrom the historical data of futures have important research value and the practical significance by changing the trading behavior into interpretable strategies.