Prototype generation on structural data using dissimilarity space representation

Calvo-Zaragoza, Jorge; Valero-Mas, Jose J.; Rico-Juan, Juan Ramón

doi:10.1007/s00521-016-2278-8

Cited by 16 publications

(8 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Calvo-Zaragoza et al (2017b) performed a classification using the nearest neighbor rule to overcome this obstacle specify in multiclass classification tasks. Calvo-Zaragoza et al (2017a) also studied the possibility of using dissimilarity space methods for mapping the initial structural representation to a statistical one, thereby allowing the use of prototype selection methods. Garcia-Pedrajas et al (2017) proposed a method for obtaining the best value by optimizing a criterion consisting of the local and global effects in the neighborhood of the prototype and they showed the advantage of the method in solving both standard and class-imbalanced problems in a large set of different problems.…”

Section: 3mentioning

confidence: 99%

Categories in emergency decision-making: prototype-based classification

Cai

Wei

Cao

2019

View full text Add to dashboard Cite

Purpose This paper aims to demonstrate how to make emergency decision when decision makers face a complex and turbulent environment that needs quite different decision-making processes from conventional ones. Traditional decision techniques cannot meet the demands of today’s social stability and security. Design/methodology/approach The main work is to develop an instance-driven classifier for the emergency categories based upon three fuzzy measures: features for an instance, solution for the instance and effect evaluation of the outcome. First, the information collected from the past emergency events is encodes into a prototype model. Second, a three-dimensional space that describes the locations and mutual distance relationships of the emergency events in different emergency prototypes is formulated. Third, for any new emergency event to be classified, the nearest emergency prototype is identified in the three-dimensional space and is classified into that category. Findings An instance-driven classifier based on prototype theory helps decision makers to describe emergency concept more clearly. The maximizing deviation model is constructed to determine the optimal relative weights of features according to the characteristics of the new instance, such that every customized feature space maximizes the influence of features shared by members of the category. Comparisons and discusses of the proposed method with other existing methods are given. Practical implications To reduce the affection to economic development, more and more countries have recognized the importance of emergency response solutions as an indispensable activity. In a new emergency instance, it is very challengeable for a decision maker to form a rational and feasible humanitarian aids scheme under the time pressure. After selecting a most suitable prototype, decision makers can learn most relevant experience and lessons in the emergency profile database and generate plan for the new instance. The proposed approach is to effectively make full use of inhomogeneous information in different types of resources and optimize resource allocation. Originality/value The combination of instances can reflect different aspects of a prototype. This feature solves the problem of insufficient learning data, which is a significant characteristic of emergency decision-making. It can be seen as a customized classification mechanism, while the previous classifiers always assume key features of a category.

show abstract

Section: 3mentioning

confidence: 99%

Categories in emergency decision-making: prototype-based classification

Cai

Wei

Cao

2019

View full text Add to dashboard Cite

show abstract

“…One possible solution by which to tackle this limitation is shown in the work by Calvo-Zaragoza et al (2017b). In this case, the initial structural data is mapped onto a statistical representation by means of the so-called dissimilarity space (DS) (Duin and Pekalska 2012) in order to then apply PG in the new representation space.…”

Section: Introductionmentioning

confidence: 99%

Prototype generation in the string space via approximate median for data reduction in nearest neighbor classification

2021

Self Cite

View full text Add to dashboard Cite

The k-nearest neighbor (kNN) rule is one of the best-known distance-based classifiers, and is usually associated with high performance and versatility as it requires only the definition of a dissimilarity measure. Nevertheless, kNN is also coupled with low-efficiency levels since, for each new query, the algorithm must carry out an exhaustive search of the training data, and this drawback is much more relevant when considering complex structural representations, such as graphs, trees or strings, owing to the cost of the dissimilarity metrics. This issue has generally been tackled through the use of data reduction (DR) techniques, which reduce the size of the reference set, but the complexity of structural data has historically limited their application in the aforementioned scenarios. A DR algorithm denominated as reduction through homogeneous clusters (RHC) has recently been adapted to string representations but as obtaining the exact median value of a set of string data is known to be computationally difficult, its authors resorted to computing the set-median value. Under the premise that a more exact median value may be beneficial in this context, we, therefore, present a new adaptation of the RHC algorithm for string data, in which an approximate median computation is carried out. The results obtained show significant improvements when compared to those of the set-median version of the algorithm, in terms of both classification performance and reduction rates.

show abstract

“…A N underlying step in machine learning and pattern recognition is the characterization of objects, where an ideally good representation ensures the building of accurate learning algorithms [1]. Three approaches have emerged to represent a real-world object [2], [3]: the structural or syntactical approach using a symbolic data structure, the statistical approach based on a feature representation, and the class models.…”

Section: Introductionmentioning

confidence: 99%