Two approaches to cluster variable quantification

Dronov, S. V.; Sazonova, Antonina S.

doi:10.3233/mas-140314

Cited by 2 publications

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Если формулировать точнее, то можно рассмотреть некую искусственную переменную, новый показатель, принимающий близкие значения на всех объектах одного и того же кластера, который аккумулирует эти существенные составляющие всех формирующих показателей. Такую переменную, следуя [6], будем называть кластерной переменной. Кластерное разбиение, следовательно, должно быть таким, чтобы условные распределения всех формирующих показателей относительно кластерной переменной оказались бы практически независимыми, ничего общего в формирующих показателях после ее фиксации не остается.…”

Section: постановка задачиunclassified

Cluster Partitions Optimization Using the Latent Class Analysis Technique

Dronov¹

2023

Известия АлтГУ

View full text Add to dashboard Cite

The paper explores the formation of a cluster partition using a system of bundles of vectors. Each object is represented by a vector with coordinates corresponding to its characteristic factors. The study focuses on the principles of changes in the total density of these bundles when objects move from one cluster to another. Based on the findings, an algorithm is proposed to reduce the total intracluster variance of the initial cluster partition. The proposed algorithm is based on a modern technique for analyzing latent classes. This technique emphasizes that the forming factors of objects within each cluster should be highly correlated. Since the number of forming factors is arbitrary, the algorithm replaces this requirement with the maximum possible proximity of object factor vectors to the average vector within each cluster. The degree of this closeness is referred to as the tightness of the cluster vector bundle. The paper also introduces a new method for quantifying clusters in the improvable clustering constructed using the proposed algorithm. A practical example of the algorithm's application to process medical data is presented. The study discusses the reasons for the dependence of the algorithm's outcome on the choice of initial cluster partition.

show abstract

Section: постановка задачиunclassified

Cluster Partitions Optimization Using the Latent Class Analysis Technique

Dronov¹

2023

Известия АлтГУ

View full text Add to dashboard Cite

show abstract

A New Algorithm for Identifying and Quantifying of Latent Classes

Dronov¹,

Shelar²

2020

Известия АлтГУ

View full text Add to dashboard Cite

Processing large amounts of data can be greatly simplified if this data is divided into approximately homogeneous groups. Splitting into such groups is the task of cluster analysis. However, the question of constructing an objective, natural partition into clusters remains open. The paper considers a modern approach to the search for such an objective cluster structure by highlighting the indicator of a common essential part from the set of characteristics that define objects (we call them the forming ones). When this indicator is fixed, the remains of the forming characteristics become independent or close to such. The resulting independent residuals are interpreted as a kind of information noise, and the latent cluster variable, the common fixed part that provides such a transformation, can be a reason for the objective integration of objects into clusters. A new algorithm for the formation of a cluster partition based on the proximity or coincidence of the values of a latent cluster variable with the simultaneous quantification of its values is proposed. The algorithm is based on the targeted search of partitions, the transition from the start one to the partition, more close to the objective. The algorithm proposed in the paper can be easily modified to the case of non-numeric categorized characteristics.

show abstract

Two approaches to cluster variable quantification

Cited by 2 publications

References 0 publications

Cluster Partitions Optimization Using the Latent Class Analysis Technique

Cluster Partitions Optimization Using the Latent Class Analysis Technique

A New Algorithm for Identifying and Quantifying of Latent Classes

Contact Info

Product

Resources

About