Abstract
Social factors such as intelligence, attitude, self-esteem (latent variables) are variables that provide relevant information to facilitate the generation of behavioral patterns or models for experts in the social area. However, such information cannot be measured directly, since they are not quantifiable. There are underlying characteristics (observable variables) to the social factors which can be measured directly to the subject of study. For this reason, a methodology based on the transformation of variables using mathematical tools is proposed. The proposed objective is to transform observable variables into hidden variables by three methods; using the arithmetic mean, Euclidean distance and exponential function. Finally, a metric based on EMD distances is applied to evaluate the similarity of the concept resulting from the three transformation methods. The EMD metric allows to evaluate the cost paid for taking one form of distribution to another, in this case the arithmetic mean and exponential function methods generate the lowest cost, that is, there is greater similarity between the distributions of the observable variables and the distribution of the resulting construct.
Original language | English |
---|---|
Title of host publication | Doctoral Symposium on Information and Communication Technologies - DSICT |
Editors | Santiago Berrezueta, Karina Abad |
Publisher | Springer Science and Business Media Deutschland GmbH |
Pages | 63-73 |
Number of pages | 11 |
ISBN (Print) | 9783030937171 |
DOIs | |
State | Published - 2022 |
Event | Doctoral Symposium on Information and Communication Technologies, DSICT 2021 - Virtual, Online Duration: 24 Nov 2021 → 26 Nov 2021 |
Publication series
Name | Lecture Notes in Electrical Engineering |
---|---|
Volume | 846 LNEE |
ISSN (Print) | 1876-1100 |
ISSN (Electronic) | 1876-1119 |
Conference
Conference | Doctoral Symposium on Information and Communication Technologies, DSICT 2021 |
---|---|
City | Virtual, Online |
Period | 24/11/21 → 26/11/21 |
Bibliographical note
Publisher Copyright:© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
Keywords
- EMD
- Latent variable
- Observable variable