| Home |
| Last parution |
| Authors |
| Topics |
| Keywords |
| Parutions |
| Videos |
| Search |
| Reading committee |
| Contact |
| Subscription |
| Article submission |
French version |
The issues from 1 to 148 are also available on line at NUMDAM and at Revues.org for the following issues |
- Results for criterions:
- Topic: Data Analysis
Results n° 1 to 8 of 172 matches
Page 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . 10 . 11 . 12 . 13 . 14 . 15 . 16 . 17 . 18 . 19 . 20 . 21 . 22 . >> Next page
| Title | When the utility, the annoyance or the comfort are recorded with ordered categories, the interaction in the multivariate context |
| Author | MAURIN Michel |
| Keywords | Category scales, Conjoint measurement, Interaction, Measurement theory, Successive intervals, Utility |
| Topics | Contingency table, Data Analysis, Measure - Measurement, Orders and preorders, Scale, Utility |
| Abstract | This note is related to ordered categorical responses coming from experimental subjects as we generally do with the «utility» in econometry, annoyance in environmental surveys, or discomfort in ergonomy. Here we are working in the measurement theory framework. The appropriate measurement method is to use the successive intervals of Adams and Messick, which are well suited for univariate conditions when responses are given in relation to one descriptive variable; here we examine the multivariate extension, the first formal properties and the possibility to explicit the interaction between variables. |
| Number | 188, Winter 2009 |
| Language | French |
| Title | A Different Approach of Multiple Correspondence Analysis (MCA) than this of Specific MCA |
| Author | MOSCHIDIS Odysseas E. |
| Keywords | Adjustement coefficient, Multiple correspondence analysis, New metric, Specific multiple correspondence analysis |
| Topics | Data Analysis, Metric, Statistics |
| Abstract | In multiple correspondence analysis, each nominal variable affects the analysis with a different amount of inertia, depending on the number of its modalities or categories. Usually in variables with many modalities - categories created infrequent (weak classes) modalities which contribute disproportionally to the inertia of the corresponding variable. Often these modalities contribute heavily to the determination of the first factorial axes and as a result this can not clearly represent the investigated problem. Specific multiple correspondence analysis deals with the problem of infrequent (weak) modalities by removing them. That is, it simply ignores them in the calculation of distances between individuals [Le Roux B., 1999, 2004]. In this paper we deal with this problem in a different manner. We keep the weak modalities in the analysis. Replacing the metric Khi2 by a new metric which also takes into account the number of modalities of each variable, a reasonable effect of the weak modalities and a balancing of all the nominal variables is achieved in the analysis. We also encounter uniformly the weak modalities, whether they derive from many or few variables, even though the most «dangerous» case is the one variables where have many modalities. Only variables of two modalities are not affected. |
| Number | 186, Summer 2009 |
| Language |  English |
| Title | Comparison of the tables of social mobility of the inquiries FQP of 1985 and 2003 by means of the tools of the graph theory: towards more continuity between occupational groups |
| Author | DALUD-VINCENT Monique |
| Keywords | Graph, Occupational groups, Social mobility, Stratification |
| Topics | Data Analysis, Graphs, Networks, Sociology |
| Abstract | The aim of this paper is to compare the tables of social mobility of the inquiries FQP of 1985 and 2003 by using the nomenclature of the PCS in 32 groups and tools of Graph Theory. The notion of connected strong component and the Réso software allow us to draw attention to the progress of the continuity between categories and certain evolutions concerning the stratification (of type «centers-suburbs») deducted. |
| Number | 185, Spring 2009 |
| Language | French |
| Title | Failure of an election is scientifically predictable! |
| Author | FANSTEN Michel |
| Keywords | Abstention, Balance of power, Clarity, Election, Homogeneous electorate, Proximity |
| Topics | Data Analysis, Decision Theory, Statistics, Voting |
| Abstract | Election behaviour, as represented in this paper, starts from a double hypothesis: given the choice of several candidates, a person will vote for the one he considers to be the closest to his or her views; within a homogeneous electorate, the factor of proximity is thus introduced according to simple statistical laws. The present article demonstrates some of the applications of mathematical formulae obtained in the case in which voters have a choice between two options or two candidates as illustrated by the recent elections in France. They basically show that voters are all the less motivated when the choices are not clearly defined, or when their favourite candidate's side is divided. In the situation where the forces of power are more or less equal, it is the greater abstention on one side that will be the determining factor for election results. |
| Number | 185, Spring 2009 |
| Language | French |
| Title | Foundations, construction and significant nodes of an implicative binary hierarchy |
| Author | LERMAN Israël-César |
| Keywords | "Significant" levels, "significant" nodes, Implicative analysis, Implicative tree, Oriented ascendant hierarchical classification |
| Topics | Classification - Clustering - Partitioning, Combinatorics, Data Analysis, Orders and preorders |
| Abstract | Implicative data analysis is a specific method of data analysis, employing asymmetrical similarities and based on a construction of an ascendant binary hierarchical classification. This method has been introduced and developed by R. Gras and collaborators. Statistical descriptive attributes (generally, boolean attributes) are organized as a binary implicative tree. In this article we reconsider in a new and systematic manner this approach. The analysis of its foundations leads us to a new axiomatic and to new constructive and combinatorial properties. We provide and motivate a new interpretation of the results obtained by this type of data analysis method.The mathematical properties of the statistical construction of the binary implicative tree are studied. Finally, we analyze in a precise and complete manner our criteria for identifying the most "significant" levels and the most "significant" nodes of the implicative binary tree. |
| Number | 184, Winter 2008 |
| Language | French |
| Title | Fionn Murtagh, "Correspondence Analysis and Data Coding with Java and R", Chapman & Hall/CRC |
| Author | ROUANET Henry, LE ROUX Brigitte |
| Keywords | None |
| Topics | Book review, Data Analysis, Lexicology, Linguistics |
| Abstract | Book review |
| Number | 181, Spring 2008 |
| Language | French |
| Title | Collinearity and linear regression analysis |
| Author | FOUCART Thierry |
| Keywords | Correlation, Instability, Principal component regression analysis, Ridge Regression, Transitivity |
| Topics | Correlation, Data Analysis, Economy - Econometrics, Linear Algebra, Regression, Statistics |
| Abstract | The linear analysis of the regression, called also more simply linear regression, is one of the most used statistical methods in applied sciences and social sciences. Its objective is double: first of all it consists in describing the relations between a variable, called explained (or dependent) variable, and several variables, called explanatory (or independent) variables. It also makes it possible to conduct forecasts of the explained variable in terms of the explanatory variables. The links between the explanatory variables exert a considerable influence on the effectiveness of the method, whatever the objective in which it is used. We expose in this paper some of the properties of these links, recently proved and published in several papers |
| Number | 173, Spring 2006 |
| Language | French |
| Title | A Gutman effect in PCA |
| Author | FLAMENT Claude, MILLAND Laurent |
| Keywords | Guttman effect, Principal composant analysis, Social representation |
| Topics | Data Analysis, Orders and preorders, Social Psychology |
| Abstract | The recourse to multidimensional methods in the data treatment allows for extracting structures of unidimensional data, when they exist, and one speaks then about Guttman effect. If the disclosure of such an effect is well-known in Correspondences Analysis, it can be also carried out in Principal Component Analysis, as presented here, while studying special properties of this effect (component of size, diagonal component, etc). The study of such a structure offers a very particular intelligibility to the results of research undertaken in the field of the social sciences, as connection with the social representation of homosexuality on the one hand and intelligence in its relationship with gender on the other hand. |
| Number | 171, Fall 2005 |
| Language | French |
Page 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . 10 . 11 . 12 . 13 . 14 . 15 . 16 . 17 . 18 . 19 . 20 . 21 . 22 . >> Next page
The entire journal is licensed under a Creative Commons license