Research On The Outlying Aspects Mining Of Data Subsets

Outlying aspects of a data subsets are attributes whose data subset is significantly different from other data subsets.It is the scientific basis for many decision-making and has important application value in real life.The outlying aspects mining of data subsets has multiple research objects,which can reflect the significant difference between one group and other groups.These outlying aspects can be used to guide researchers and managers in making and optimizing relevant decisions.At present,the main research objects of the various methods of outlying aspects mining are single,and they do not involve many research objects.Moreover,the existing method cannot effectively excavate the outlying aspects of the data subsets and it is difficult to meet the needs of analyzing the attribute differences between groups in reality.Aiming at these problems in the research,this paper presents an outlying aspects mining problem based on the data subsets of multiple research obj ects.The outlying aspects mining problem of the proposed data subsets is formally defined,and the Group Outlying Aspects Mining(GOAM)algorithm is designed.Experimental results show that compared with the traditional outlying aspects mining algorithm,the algorithm can identify the significant attributes of the data subsets.Then it effectively solves the problem of outlying aspects mining in real life about multiple research objects.In order to further verify and illustrate the usability and practicability of the GOAM algorithm proposed in this paper,the GOAM algorithm is applied to two real datasets,12 groups of outlying aspects mining experiments on the subset of data are carried out,and the remarkable attribute results obtained by these mining are visualized and analyzed.The results of these analyses can be used as a scientific basis for practitioners to optimize management decisions and develop strategic plans.