Omni-Directional Time Evolution Analysis Based on Multidimensional Financial Big Data
Shuhua Wang · 2019
The omni-directional time evolution analysis method is an important method for learning the financial big data stream with sample drift. Aiming at the shortcomings of the traditional time evolution analysis method, the human memory curve is introduced into the time evolution analysis, and an Omnidirectional Time Evolution Analysis Method (OTEAM) is proposed, which uses the Ebbinghaus curve. The system's memory curve is designed and used selectivity to simulate the human "memory" mechanism. Compared with four typical analysis methods, the results show that the OTEAM algorithm has high classification accuracy and strong overall adaptability to sample drift, especially for repetitive sample drift and complex sample drift in practical applications. Not only can it quickly adapt to new sample changes, but it can also effectively resist the effects of random sample fluctuations.