Cross-Validation Approach to Evaluate Clustering Algorithms: An Experimental Study Using Multi-Label Datasets

Tarekegn, Adane Nega; Michalak, Krzysztof; Giacobini, Mario

doi:10.1007/s42979-020-00283-z

Cited by 12 publications

(6 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Each dataset contains one class variable that represents the target outcome. The dataset has been used in various applications and research papers, such as [17,18].…”

Section: Datasetsmentioning

confidence: 99%

Evolutionary Machine Learning for Intelligent Health Risk Prediction

Tarekegn,

Alemu

2024

Preprint

View full text Add to dashboard Cite

Evolutionary Machine Learning (EML) has demonstrated its effectiveness in modeling various real-world application domains, including the dynamics of diseases and the creation of predictive models across diverse fields. However, the challenge of data imbalance adversely impacts the performance of EML models. While strategies involving misclassification costs or cost-sensitive learning exist to address this issue, their integration into EML has yet to be thoroughly explored. This paper aims to fill this gap by investigating the application of cost-sensitive EML for intelligent health risk prediction. More specifically, this paper delves into the utilization of cost-sensitive genetic programming (CSGP) and cost-sensitive genetic algorithms (CSGA), which are among the most widely used algorithms in the EML category. The study entails training CSGP and CSGA to predict the early onset of three chronic conditions (hypertension, diabetes, and fatty acid-related disorders) through the development of an EML model. Subsequently, the model's performance is evaluated through standard fitness functions such as sensitivity, specificity, and F1-score. From the experimental results, EML models demonstrate promising results in identifying the occurrence of chronic diseases. Further optimization of the cost-sensitive EML model holds the potential for enhanced outcomes in modeling medical problems.

show abstract

“…Each dataset contains one class variable that represents the target outcome. The dataset has been used in various applications and research papers, such as [17,18].…”

Section: Datasetsmentioning

confidence: 99%

Evolutionary Machine Learning for Intelligent Health Risk Prediction

Tarekegn,

Alemu

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Clustering quality evaluation also provides additional insight into the underlying structure of data, which is especially important [ 28 ] in real-world applications. Quality assessment techniques can be divided into internal and external clustering validity methods [ 29 ]. Internal methods e.g., the silhouette coefficient, Dunn index, Davies–Bouldin index, and Calinski–Harabasz index, evaluate clustering quality based on the data distribution and inter-cluster relationships.…”

Section: State Of the Artmentioning

confidence: 99%

Comparative Analysis of the Clustering Quality in Self-Organizing Maps for Human Posture Classification

Ekemeyong Awong,

Zielinska

2023

Sensors

View full text Add to dashboard Cite

The objective of this article is to develop a methodology for selecting the appropriate number of clusters to group and identify human postures using neural networks with unsupervised self-organizing maps. Although unsupervised clustering algorithms have proven effective in recognizing human postures, many works are limited to testing which data are correctly or incorrectly recognized. They often neglect the task of selecting the appropriate number of groups (where the number of clusters corresponds to the number of output neurons, i.e., the number of postures) using clustering quality assessments. The use of quality scores to determine the number of clusters frees the expert to make subjective decisions about the number of postures, enabling the use of unsupervised learning. Due to high dimensionality and data variability, expert decisions (referred to as data labeling) can be difficult and time-consuming. In our case, there is no manual labeling step. We introduce a new clustering quality score: the discriminant score (DS). We describe the process of selecting the most suitable number of postures using human activity records captured by RGB-D cameras. Comparative studies on the usefulness of popular clustering quality scores—such as the silhouette coefficient, Dunn index, Calinski–Harabasz index, Davies–Bouldin index, and DS—for posture classification tasks are presented, along with graphical illustrations of the results produced by DS. The findings show that DS offers good quality in posture recognition, effectively following postural transitions and similarities.

show abstract

“…Emerging machine learning algorithms, particularly clustering algorithms, have found applications in data grouping studies across various fields. These algorithms have proven to be scientific, objective, and efficient in classification studies [38]. They do not require prior knowledge of data object categories or labels and group them based on similarity or distance.…”

Section: Literature Reviewmentioning

confidence: 99%

Identification of Inefficient Urban Land for Urban Regeneration Considering Land Use Differentiation

Jin,

Huang,

Wang

et al. 2023

Land

View full text Add to dashboard Cite

Accurately identifying inefficient urban land is essential for urban regeneration and mining underutilized assets. Previous studies have primarily focused on examining the overall efficiency of land use without adequately considering the heterogeneity of urban land use types and comprehensive characteristics of urban quality. As a result, the spatial accuracy and precision of research findings have been relatively low. To address this gap, we developed a comprehensive method to identify inefficient urban lands for residential, commercial, and industrial use. The method integrated multi-source geographic data to quantitatively characterize the efficiency of different land use types considering six key dimensions, including building attribute, urban service, transportation condition, environmental quality, business performance, and production efficiency, utilized principal component analysis to reduce the multicollinearity and the dimensionality of the data, and identified land clusters with similar features that were inefficiently used by means of hierarchical clustering. By applying the method to Changsha, China, we validated its effectiveness. The results demonstrate that the method can accurately identify inefficient residential, commercial, and industrial land, with kappa coefficients of 0.71, 0.77, and 0.68, respectively. The identification results reveal the spatial distribution patterns of different types of inefficient land. Inefficient residential land is concentrated towards the city center, particularly in central areas. Inefficient commercial land is relatively evenly distributed, mainly outside the core commercial regions. Inefficient industrial land clusters towards the periphery, forming several agglomeration areas centered around industrial parks. By precisely identifying inefficient urban land and focusing on the key influencing factors, the proposed method enables the site selection of urban regeneration, site redevelopment evaluation, and optimization of urban resources.

show abstract

Cross-Validation Approach to Evaluate Clustering Algorithms: An Experimental Study Using Multi-Label Datasets

Cited by 12 publications

References 37 publications

Evolutionary Machine Learning for Intelligent Health Risk Prediction

Evolutionary Machine Learning for Intelligent Health Risk Prediction

Comparative Analysis of the Clustering Quality in Self-Organizing Maps for Human Posture Classification

Identification of Inefficient Urban Land for Urban Regeneration Considering Land Use Differentiation

Contact Info

Product

Resources

About