01755nas a2200253   4500000000100000000000100001008004100002260001200043653001000055653002100065653001900086653002000105653002600125653002700151100002700178700003100205700003700236245010700273856005800380300000900438490001300447520102700460022001401487       9998                            d  c11/202410aForex10aMachine Learning10aMicro Services10aReduction Model10aSoftware Architecture10aSupport Vector Machine1 aKevin Gordillo-Orjuela1 aPaulo Alonso Gaona-García1 aCarlos Enrique Montenegro-Marín00aDesign of a Machine Learning-Based Platform for Currency Market Prediction: A Fundamental Design Model  uhttps://www.ijimai.org/journal/bibcite/reference/3511  a1-110 vIn press3 aPrediction models in foreign exchange markets have been very popular in recent years, and in particular, through the use of techniques based on Machine Learning. This growth has made it possible to train several techniques that increasingly allow us to improve predictions according to the criteria that each algorithm supports and can cover. However, the development of these models and their deployment within computer platforms is a complex task, given the variety of approaches that each researcher uses based on the training process and therefore by definition of the model, which leads to the consumption of high computing resources for its training, as well as various processes for its deployment. For this reason, the following article focuses on designing a technological platform oriented to micro services, which minimizes the consumption of resources and facilitates the integration of various techniques and the analysis of various criteria, which improves their analysis and validation in a Web environment.  a1989-1660