China 
Africa 
Explore chapters and articles related to this topic
The Standardization of the Measurement of the Effects of Optimization on Management Practices
Published in Walter Amedzro St-Hilaire, Value-Based Management in an Open Economy, 2023
After the construction of the performance measurement index, a characterization of these Banking sector companies according to the relevant variables is carried out. For this purpose, Multiple Factor Analysis is used. This technique is applied in the case of multiple tables where the same group of individuals is described by several groups of variables. It is based on principal component analysis (PCA) when the variables in each group are quantitative and on multiple correspondence analysis (MCA) when the variables in each group are qualitative. Since in our case the variables are all quantitative, it is the PCA that is at the heart of the Multiple Factor Analysis. Thus, the data are presented in the form of multiple tables, i.e., in the form of an I observation table where the variables K are split into groups J. PCA is a method of factor analysis which consists of searching for reduced-dimensional subspaces which best fit the cloud of points-individuals (the rows of the initial data table representing the branches) and the cloud of points-variables (the columns of the initial data table representing the variables).
Identifying driving behaviour profiles by using multiple correspondence analysis and cluster analysis
Published in Gianluca Dell’Acqua, Fred Wegman, Transport Infrastructure and Systems, 2017
D.S. Usami, L. Persia, M. Picardi, M.R. Saporito, I. Corazziari
Multiple Correspondence Analysis (MCA) is a factorial analysis method allowing summarizing a set of categorical variables by one or few quantitative and not correlated (orthogonal) ones. The main aim of MCA is to describe the association among the original variables, and to measure the underlying latent factor not directly observable, represented by the set of observable variables.
Mapping the Cybersecurity Research: A Scientometric Analysis of Indian Publications
Published in Journal of Computer Information Systems, 2023
B. Elango, S. Matilda, M. Martina Jose Mary, M. Arul Pugazhendhi
As a result of the investigation, multiple correspondence analysis was used to build conceptual structure maps of the author keywords (2011–2017 and 2018–2020) and the resulting clusters were shown in two-dimensional maps (see Figure 12): thus, two clusters were formed in the two phases. Multiple Correspondence Analysis (MCA) is a commonly used technique for the analyzing categorical data with the purpose of reducing a large collection of data into smaller sets of components. The closer the dots on the graph represent each keyword, the more similar the distribution of keywords is, meaning that they co-occur more frequently in the publications. Keywords near the center are of great interest to the research community, whereas keywords on the periphery are of low degree of relevance to other research topics.47,48
New data-driven approach to generate typologies of road segments
Published in Transportmetrica A: Transport Science, 2023
Asad Yarahmadi, Catherine Morency, Martin Trepanier
The last feature engineering technique that was performed was Dimensionality Reduction. This technique is mainly used to find and safely drop items with the lowest variance values compared to other variables. Generally, there are three techniques: Principal Component Analysis (PCA) is for continuous values, Multiple Correspondence Analysis (MCA) is suitable for categorical values, and Factor Analysis Mixed Data (FAMD) is used for the mixed data type. This research uses PCA and FAMD to prepare data for clustering approach one and two, respectively. The PCA generates a covariance matrix of all features and then calculates their eigenvectors and eigenvalues, showing their magnitude of variance. Next, the number of components needed to describe should be determined according to the eigenvalues. To do that, the cumulative explained variance ratio is utilised, and Figure 3 illustrates that 95% of components are needed to explain variance. Another technique is the FAMD. Implementing the FAMD on mixed data types is necessary because it balances the influences of categorical and numerical values on the analysis. To do that, first, all numerical values were scaled to variance, and the MCA was utilised to scale categorical values. The MCA is composed of two following steps. First, a one-hot encoded version is computed, and then a corresponding analysis is executed to generate a contingency table by analysing dependencies between categorical variables.
Artificial Intelligence and Human Resources Management: A Bibliometric Analysis
Published in Applied Artificial Intelligence, 2022
P.R. Palos-Sánchez, P. Baena-Luna, A. Badicu, J.C. Infante-Moro
Conceptual structure: It refers to what the science is about, the main themes, and trends. Specifically, multiple correspondence analysis (MCA) helps analyze categorical data to reduce large sets of variables into smaller sets to synthesize the information in the data (Mori et al. 2014). To do this, the data are compressed into a low-dimensional space to form a dimensional or three-dimensional graph that uses planar distance to reflect the similarity between keywords.