A Grouping Method for Categorical Attributes Having Very Large Number of Values

Author(s):  
Marc Boullé
Algorithms ◽  
2021 ◽  
Vol 14 (6) ◽  
pp. 184
Author(s):  
Xia Que ◽  
Siyuan Jiang ◽  
Jiaoyun Yang ◽  
Ning An

Many mixed datasets with both numerical and categorical attributes have been collected in various fields, including medicine, biology, etc. Designing appropriate similarity measurements plays an important role in clustering these datasets. Many traditional measurements treat various attributes equally when measuring the similarity. However, different attributes may contribute differently as the amount of information they contained could vary a lot. In this paper, we propose a similarity measurement with entropy-based weighting for clustering mixed datasets. The numerical data are first transformed into categorical data by an automatic categorization technique. Then, an entropy-based weighting strategy is applied to denote the different importances of various attributes. We incorporate the proposed measurement into an iterative clustering algorithm, and extensive experiments show that this algorithm outperforms OCIL and K-Prototype methods with 2.13% and 4.28% improvements, respectively, in terms of accuracy on six mixed datasets from UCI.


2021 ◽  
Vol 27 (1) ◽  
pp. 41-62
Author(s):  
Konstantin V. KRINICHANSKII

Subject. This article examines the non-financial sector debt ratio relative to the GDP of emerging and developed market economies. Objectives. The article aims to find out what institutional units show debt growth or reduction, what causes and conditions prevent debt decline or lead to its growth in different countries and sectors, and highlight the foundations of public policy in this area. Methods. For the study, I used a cross-country comparative analysis, grouping method, and graphical and trend analyses. The study covers 43 market economies, including 26 developed and 17 emerging ones. The time period is from Q4 2001 to Q4 2019. Results. The article identifies and describes the structural debt changes that have taken place since 2008, which include a reduction in private sector leverage and rising public sector debt in developed market economies, and accelerated growth in the non-financial corporations and households' debt in emerging market economies. Conclusions and Relevance. Given the different conditions of access to the capital market and the institutional differences between developed and emerging market economies, different approaches to debt management are needed. The identified trends are important to develop non-financial sector debt management policies, including both fiscal and monetary policies.


2016 ◽  
Vol 211 ◽  
pp. 191-201 ◽  
Author(s):  
Qiuling Hou ◽  
Ling Zhen ◽  
Naiyang Deng ◽  
Ling Jing

Author(s):  
Diêgo Lima Crispim ◽  
Lindemberg Lima Fernandes ◽  
Roberta Luiza de Oliveira Albuquerque

Indicators are important tools to guide and assist decision-makers. They are also important to get to know the scenario of a given place and monitor its development. This study aimed to analyze the behavior of the municipalities of Marajó-PA through indicators that cover social, economic, housing and sanitation, using a statistical technique of multivariate analysis to group these into a small number of homogeneous groups. In order to choose the indicators, we carried out a checklist of national, regional and local academic papers dealing with sustainability. Then, the indicators were standardized according to the different units and scales of measurement, not influencing the result and presenting similar weights in the calculation of the similarity coefficient. The measure of dissimilarity used was the euclidean distance and for the composition of the groupings the Ward and k-Means methods were applied. The result obtained using Ward’s hierarchical grouping method enabled the reduction of the numbers of municipalities to a number of 4 probable groups with similar attributes within the group and distinct among the others. It also presented a cofenetic correlation coefficient (CCC) of (r = 0.81), indicating a good degree of fit between the dendrogram and the dissimilarity matrix. The results indicated that the formation of the clusters and the municipalities integrated in them presented similarity both in the hierarchical and non-hierarchical methods. In the k-means method it was found that almost all municipalities that make territorial division remained within the same group.


Author(s):  
Barath Kumar R ◽  
Stalin Alex

The immovable nature of passing on packages through multi -bounce middle hubs is a critical problem in the versatile impromptu organizations (MANETs). The disseminated versatile hubs set up associations with structure the MANET, which may incorporate childish and getting into mischief hubs. Suggestion based trust the board is proposed in the creating as a system to evaluate through the acting up hubs while looking for a bundle conveyance course. Nonetheless, building a trust model that embraces suggestions by different hubs in the organization is a difficult issue because of the danger of deceptive proposals like reviling, voting form stuffing, and conspiracy. we examines the issues identified with assaults presented by getting rowdy hubs while proliferating suggestions in the current trust models. We propose a suggestion based trust model with a safeguard plot, which uses grouping method to progressively sift through assaults identified with exploitative proposals between certain time dependent on number of collaborations, similarity of data and closeness between the hubs. We evaluate the trust degree as two cases like direct and indirect trust values between neighboring nodes from source. To form a clustering routing network from similar trust values from S to D.The model is experimentally tried under a few portable and detached geographies in which hubs experience changes in their local prompting regular course changes. The observational investigation shows heartiness and exactness of the trust model in a dynamic MANET climate.


2020 ◽  
Vol 4 (3) ◽  
Author(s):  
Yu Zhu ◽  
Lu Shi

Objective: To analyze the clinical treatment effect of traditional Chinese medicine five-color therapy on chronic urticaria in children. Methods: The income data target of this article is 80 children with chronic urticaria. The grouping method is a randomized method with 40 children in each group. The experimental group was treated with five-color treatment of traditional Chinese medicine, and the control group was treated with western medicine. The incidence, treatment and recurrence of adverse reactions in children with chronic urticaria were compared between the two groups. Results: Showed total effective rate of children with chronic urticaria in the experimental group was compared with the control group, P<0.05, the data showed statistical significance. Conclusion: Stated use of TCM five-color therapy in the treatment of children with chronic urticaria can significantly improve safety.


Sign in / Sign up

Export Citation Format

Share Document