Feature selection is an important data-preprocessing technique in classification problems such as bioinformatics and signal processing. Generally, there are some situations where a user interested not only maximizing the performance but also minimizing cost that may be associated with features. This kind of problem called cost-based feature selection. However, most existing approaches treat this task single-objective optimization problem. paper presents first study multi-objective particle...

Evolutionary feature selection (FS) methods face the challenge of "curse dimensionality" when dealing with high-dimensional data. Focusing on this challenge, article studies a variable-size cooperative coevolutionary particle swarm optimization algorithm (VS-CCPSO) for FS. The proposed employs idea "divide and conquer" in approach, but several new developed problem-guided operators/strategies make it more suitable FS problems. First, space division strategy based importance is presented,...

The "curse of dimensionality" and the high computational cost have still limited application evolutionary algorithm in high-dimensional feature selection (FS) problems. This article proposes a new three-phase hybrid FS based on correlation-guided clustering particle swarm optimization (PSO) (HFS-C-P) to tackle above two problems at same time. To this end, three kinds methods are effectively integrated into proposed their respective advantages. In first second phases, filter method...

Feature selection (FS) is an important data processing technique in the field of machine learning. There have been various FS methods, but all assume that cost associated with a feature precise, which restricts their real applications. Focusing on problem fuzzy cost, multiobjective method particle swarm optimization, called PSOMOFS, studied this article. The proposed develops dominance relationship to compare goodness candidate particles and defines crowding distance measure prune elitist...

Various real-world multiobjective optimization problems are dynamic, requiring evolutionary algorithms (EAs) to be able rapidly track the moving Pareto front of an problem once environmental change occurs. To this end, several methods have been developed predict new location set (PS) so that population can reinitialized around predicted location. In paper, we present a multidirectional prediction strategy enhance performance EAs in solving dynamic (DMOP). more accurately PS, is clustered...

Dynamic interval multiobjective optimization problems (DI-MOPs) are very common in real-world applications. However, there few evolutionary algorithms (EAs) that suitable for tackling DI-MOPs up to date. A framework of dynamic cooperative co-evolutionary based on the similarity is presented this paper handle DI-MOPs. In framework, a strategy decomposing decision variables first proposed, through which all divided into two groups according between each variable and parameters. Following that,...

Feature selection is an important data preprocessing technique in multi-label classification. Although a large number of studies have been proposed to tackle feature problem, there are few cases for data. This paper algorithm using improved multi-objective particle swarm optimization (PSO), with the purpose searching Pareto set non-dominated solutions (feature subsets). Two new operators employed improve performance PSO-based algorithm. One operator adaptive uniform mutation action range...

Various real-world applications can be classified as expensive multimodal optimization problems. When surrogate-assisted evolutionary algorithms (SAEAs) are employed to tackle these problems, they not only face a contradiction between the precision of surrogate models and cost individual evaluations but also have difficulty that problem modalities hard match. To address this issue, article studies dual-surrogate-assisted cooperative particle swarm algorithm seek multiple optimal solutions. A...

Index plays an essential role in modern database engines to accelerate the query processing. The new paradigm of "learned index" has significantly changed way designing index structures DBMS. key insight is that indexes could be regarded as learned models predict position a lookup dataset. While such studies show promising results both time and size, they cannot efficiently support update operations. Although recent have proposed some preliminary approaches update, are at cost scarifying...