Research On Feature Selection Method Based On Intelligent Optimization Algorithm Of Whale Swarm

In today’s era of increasingly mature big data and intelligent development.A large number of data will be generated in production and life.The analysis of these data and the application of the analysis results to production and life can improve the efficiency and quality of production in all walks of life.This is where the data analysis method feature selection is used.However,the traditional feature selection method is no longer suitable for the analysis of the huge data set in terms of the accuracy rate of classification and the number of selected features,so it needs to be optimized.Therefore,it is found in the research that swarm intelligent algorithm can well find the optimal subset in feature selection by using its iterative optimization features,so it is combined.However,at present,the traditional group optimization algorithm has some problems such as low precision and slow convergence speed in some specific fields,so the application effect in the feature is improved but not obvious.Therefore,the traditional swarm intelligence algorithm should be optimized first,and then introduced into feature selection.Because there are many swarm intelligence algorithms,I choose whale swarm intelligence algorithm here,because its model is concise and its parameters are few,and its performance is as good as classical algorithms such as fish algorithm and wolf colony algorithm.But there are many shortcomings.Therefore,this paper firstly optimizes and improves the traditional whale swarm algorithm,and then meets the actual needs of feature selection to improve the accuracy rate of classification and reduce the number of features.Feature selection is of high research significance at present,as can be seen from many current classification algorithms,and it is also an indispensable link in artificial intelligence,which is used to eliminate redundant data and noisy features from huge data sets formed in production and life.Therefore,there are two main points in this paper.First,the whale swarm algorithm in the swarm intelligence algorithm is optimized and improved.The main idea of improvement is to focus on the optimization of the whole whale swarm intelligence algorithm in the two stages of exploration and development,so that whales can adjust the search mechanism by introducing random factors in the search for prey stage in the exploration stage.Random factors are introduced to improve the fast convergence rate of the algorithm in the encircling prey stage of the development stage,and factor Q is introduced in the spiral stage to improve the speed of individuals approaching the optimal solution,that is,to improve the optimization accuracy.Therefore,these optimization ideas can solve the problems of low precision and slow convergence of the algorithm,and the above strategies can improve the accuracy and robustness of the algorithm.Secondly,the optimized whale swarm algorithm is applied to the feature selection algorithm.And through the experiment,the classification and selection results are obviously improved.Therefore,it is very valuable to introduce intelligent algorithm in feature extraction and simplify data preprocessing process.Through 9 reference functions,which are composed of single-peak function and multi-peak function,the optimized whale swarm algorithm and others’ improved algorithm are simulated and analyzed.Secondly,in the application,the application feature selection algorithm and the original feature selection algorithm were compared on 7 standard complex data sets.The analysis of experimental results shows that the optimized algorithm has good global search ability and development potential both on the standard data set and on practical engineering problems.The effectiveness of algorithm optimization and application is verified by simulation.