Research output: Contribution to journal › Article › peer-review
Application of Random Forest in Choosing a Method of Recovering the Age of Social Network Users. / Корепанова, Анастасия Андреевна; Абрамов, Максим Викторович.
In: Scientific and Technical Information Processing, Vol. 49, No. 5, 01.12.2022, p. 317–324.Research output: Contribution to journal › Article › peer-review
}
TY - JOUR
T1 - Application of Random Forest in Choosing a Method of Recovering the Age of Social Network Users
AU - Корепанова, Анастасия Андреевна
AU - Абрамов, Максим Викторович
PY - 2022/12/1
Y1 - 2022/12/1
N2 - Abstract: This article is devoted to the problem of recovering the ages of social network users by using machine learning to combine the methods suggested in this article. Methods based on analyzing the user profile information about education, subscription, and information about the education of the user’s friends are considered. All of these methods can be used individually for samples of users with certain characteristics. To increase the proportion of users whose age can be recovered, a classification model was built for choosing the best age inference method for each. Two other age inference algorithms were tested as well, namely, ranking and score averaging. As a result, the first approach produced the best results on the test sample. The theoretical significance of this work consists in proposing a method of combining age inference algorithms, which increases the applicability and accuracy of individual algorithms. The study results can be applied in many spheres of analyzing user profiles in social network studies.
AB - Abstract: This article is devoted to the problem of recovering the ages of social network users by using machine learning to combine the methods suggested in this article. Methods based on analyzing the user profile information about education, subscription, and information about the education of the user’s friends are considered. All of these methods can be used individually for samples of users with certain characteristics. To increase the proportion of users whose age can be recovered, a classification model was built for choosing the best age inference method for each. Two other age inference algorithms were tested as well, namely, ranking and score averaging. As a result, the first approach produced the best results on the test sample. The theoretical significance of this work consists in proposing a method of combining age inference algorithms, which increases the applicability and accuracy of individual algorithms. The study results can be applied in many spheres of analyzing user profiles in social network studies.
KW - attribute inference
KW - machine learning
KW - social engineering attacks
KW - social media analysis
KW - sociocomputing
UR - https://www.mendeley.com/catalogue/84a03436-ebb3-3e99-9dc3-61442a8f422c/
U2 - 10.3103/s0147688222050057
DO - 10.3103/s0147688222050057
M3 - Article
VL - 49
SP - 317
EP - 324
JO - Scientific and Technical Information Processing
JF - Scientific and Technical Information Processing
SN - 0147-6882
IS - 5
ER -
ID: 113494877