Technical Analysis on Financial Time Series Data based on Map-Reduce Programming Model: A Case Study

Abstract

Technical analysis is a widely used method for forecasting the price direction on the financial time series data. This method requires the use of different number and types of analysis algorithms (technical indicators) together. Although these algorithms show successful performance on small-scale financial time series data, significant performance decreases are detected when the size of data increased. On the large-scale financial time series data, it is necessary to implement these algorithms based on the Map-Reduce programming model and examine the performance of the algorithms which are implemented based on this model comparatively. For this purpose, seven different indicators are studied within the scope of this study, new versions of these indicators are implemented using Map-Reduce parallel data processing model and performance comparisons are made with these algorithms. As a result of these comparisons on single-node and multi-node, significant performance gains have been obtained using Map-Reduce programming model.