Lead selecting method based on improved upsemoa algorithm

Abstract

The invention provides a lead selecting method based on an improved upsemoa algorithm. The method comprises the following steps of (1), acquiring a population, acquiring input parameters which comprises a population minimal size miSize, a father-generation basis vector number burstSize, and an individual length L, randomly generating miSize individuals with length of L in a given searching space for forming the population P; (2), performing initial determining, determining whether the population P is a first generation, if not, combining the output filial generation population C into the population P, and initializing a non-inferior-solution set T; (3), selecting the set, choosing the non-inferior-solution set Te from the population P; (4), performing crowdedness sequencing, performing crowdedness comparison sequencing on the non-inferior-solution set T; (5), performing choosing and combination; (6), choosing the father-generation basis vector; (7), selecting the filial generation population; and (8), ending determining. The lead selecting method realizes a time length which is far shorter than that of other lead selecting methods based on multi-target evolution under a precondition that classification accuracy is not lower than other lead selecting methods.

Description

technical field [0001] The invention relates to a lead selection method based on an improved upsemoa algorithm. Background technique [0002] The structure of the brain is complex and delicate, and it is an important field of life science research. It is related to human life activities and spiritual activities. As European and American scientific research powers have launched their own brain programs, my country has also formulated a brain program called "Brain Science and Brain-Inspired Intelligence" in the 15-year plan for 2016-2030. [0003] People's research on the brain can be traced back to Cajal's creation of the "neuron theory", which has a history of more than 100 years. Despite this, we still know very little about the mysteries of the brain. At present, the methods for brain experiments include electroencephalogram (Electroencephalogram, EEG) technology with high temporal resolution and functional magnetic resonance imaging (Functional Magnetic Resonance Imagi...

AI Technical Summary

Problems solved by technology

In addition, the lead selection method based on wrapping only pursues a single target. Even if the single target is converted into multiple targets through the weight coefficient, it is difficult to consider the best of both worlds for the two targets.

Although Kee uses the lead selection method based on the NSGA-II algorithm to make a trade-off between the number of leads and the classification accuracy, the lead selection method requires a high time cost

Images