TY - JOUR
T1 - Psychological attachment style determination using a word space model
AU - Calvo, Hiram
N1 - Publisher Copyright:
© 2018 Instituto Politecnico Nacional. All rights reserved.
PY - 2018
Y1 - 2018
N2 - In this work we experiment with the hypothesis that words a subject uses and their psychological attachment style (as defined by Bartholomew and Horowitz), can be related. In order to verify this hypothesis, we identified characteristic patterns for each style of attachment (secure, fearful, dismissing, preoccupied) by mapping words into a word space model on a series of autobiographic texts written by a set of 202 participants. Additionally, a psychological instrument (questionnaire) was applied to these same participants to measure their attachment style. A Support Vector Machine was trained, and we found that attachment style could be predicted from text within a range of 64% to 85% for different attachment styles.
AB - In this work we experiment with the hypothesis that words a subject uses and their psychological attachment style (as defined by Bartholomew and Horowitz), can be related. In order to verify this hypothesis, we identified characteristic patterns for each style of attachment (secure, fearful, dismissing, preoccupied) by mapping words into a word space model on a series of autobiographic texts written by a set of 202 participants. Additionally, a psychological instrument (questionnaire) was applied to these same participants to measure their attachment style. A Support Vector Machine was trained, and we found that attachment style could be predicted from text within a range of 64% to 85% for different attachment styles.
KW - Autobiography
KW - Psychological attachment
KW - Support vector machines
KW - Text classification
UR - http://www.scopus.com/inward/record.url?scp=85069557636&partnerID=8YFLogxK
U2 - 10.13053/CyS-22-4-3064
DO - 10.13053/CyS-22-4-3064
M3 - Artículo
AN - SCOPUS:85069557636
SN - 1405-5546
VL - 22
SP - 1269
EP - 1276
JO - Computacion y Sistemas
JF - Computacion y Sistemas
IS - 4
ER -