Group Labeling Methodology Using Distance-based Data Grouping Algorithms

Abstract

Clustering algorithms are often used to form groups based on the similarity of their members. In this context, understanding a group is just as important as its composition. Identifying, or labeling groups can assist with their interpretation and, consequently, guide decision-making efforts by taking into account the features from each group. Interpreting groups can be beneficial when it is necessary to know what makes an element a part of a given group, what are the main features of a group, and what are the differences and similarities among them. This work describes a method for finding relevant features and generate labels for the elements of each group, uniquely identifying them. This way, our approach solves the problem of finding relevant definitions that can identify groups. The proposed method transforms the standard output of an unsupervised distance-based clustering algorithm into a Pertinence Degree (GP), where each element of the database receives a GP concerning each formed group. The elements with their GPs are used to formulate ranges of values for their attributes. Such ranges can identify the groups uniquely. The labels produced by this approach averaged 94.83% of correct answers for the analyzed databases, allowing a natural interpretation of the generated definitions.

Method for Inferring the Optimal Number of Clusters with Subsequent Automatic Data Labeling based on Standard Deviation

“…The Table XIV displays the result of labeling the method of [20] for the Iris dataset. It was observed that the result was generated after 568 iterations of the method based on degrees of membership, which corresponds to a very high computational cost, in addition to having 12 database elements that could not be labeled.…”

“…This model, proposed in this research, generated a maximum of six iterations to form labels, corroborating and reducing the computational effort. No criteria were used to infer the optimal number of groups in [20], so the author used K=3 for clustering. In this proposed model, it was observed that the computational cost spent on forming the labels was extremely low, favoring a minimum number of iterations, compared to the [20] model, which presented a lower rate than this research model.…”

“…No criteria were used to infer the optimal number of groups in [20], so the author used K=3 for clustering. In this proposed model, it was observed that the computational cost spent on forming the labels was extremely low, favoring a minimum number of iterations, compared to the [20] model, which presented a lower rate than this research model. Using the proposed method for inferring the optimal amount of clusters favored the obtained result.…”