Context-aware Personalized Decision Support based on User Digital
Life Model
Alexander Smirnov
a
and Tatiana Levashova
b
St. Petersburg Federal Research Center of the Russian Academy of Sciences, St. Petersburg, Russia
Keywords: Decision Support, Personalized Recommendations, Personality, User Ontology, Digital Traces, User Digital
Life Model.
Abstract: Digital traces is a source of information about the users and their actions while the online activities. A
structured part of this source, which represents information related to the decision-making process of a user,
is proposed to be formalized in the form of user digital life model. The decision support system addresses this
model for information to recognize user types and recommend personalized decisions. A user type is
characterized with common personality traits of the users as decision makers and common decision-making
behaviours of these users as consumers. A user ontology represents a priori knowledge on the user types and
supports the user classification into them. The paper considers kinds of factors influencing decision-making
styles and consequently personality traits of decision makers as well as behaviour variables determining
decision-making behaviour. The user digital life model provides information to score these factors and
instantiate the variables. A decision support scenario is described and its application to a search problem is
demonstrated.
1 INTRODUCTION
Personalized decision support based on information
from user digital traces has recently gained popularity
as a result of the success of the efforts on personality
prediction from these traces (Stachl, Pargent, et al.,
2020). Digital traces are beneficial for obtaining
users’ personality traits without burdensome
questionnaires. Decision support systems (DSSs) and
recommendation systems add the personality
information in the processes of decision support and
recommendation.
One of the problems of using information from
digital traces is their weakly structured and poorly
curated content that complicates its analysis in a
context-aware meaningful way (Breiter & Hepp,
2018). Information from the digital traces represented
in a well-structured way would simplify its analysis
to predict personality.
Context-aware integration of information
describing personality traits and online behaviour that
the user manifest while decision-making to predict or
a
http://orcid.org/0000-0001-8364-073X
b
http://orcid.org/0000-0002-1962-7044
recognize personality would increase the efficiency
of DSSs.
The paper proposes a conceptual framework of
personalized decision support based on user digital
life model, in which such a model systematizes the
content of digital traces and represents information
related to the decision-making process of a user. This
model is used as an information source to recognize
user types and recommend personalized decisions. A
user type groups users with common personality traits
of them as decision makers and common decision-
making behaviours as consumers. A user ontology
represents a priori knowledge on the user types and
supports the user classification into these types. The
types are context-sensitive, i.e. the same user in
different contexts can be classified into different user
types. A DSS that implements the conceptual
framework infers the user type and recommends a
decision based on the knowledge about the kinds of
decisions customary for the users of this type.
The rest of paper is as follows. Section 2 outlines
related research. Section 3 introduces a conceptual
framework of personalized decision support based on
Smirnov, A. and Levashova, T.
Context-aware Personalized Decision Support based on User Digital Life Model.
DOI: 10.5220/0011526900003323
In Proceedings of the 6th International Conference on Computer-Human Interaction Research and Applications (CHIRA 2022), pages 129-136
ISBN: 978-989-758-609-5; ISSN: 2184-3244
Copyright
c
2022 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
129
user digital life model. Section 4 presents the user
ontology and provides ideas of user classification.
Section 5 describes a scenario of personalized
decision support according to the conceptual
framework. Section 6 demonstrates the scenario
applied to a book search problem. Section VII
discusses main concluding remarks.
2 RELATED RESEARCH
Many studies have concluded that personalities
influence human decision making process and
interests (Rentfrow & Gosling, 2003).
The most commonly, research on personality
prediction from user digital traces analyses traces
from social media platforms. Such approaches
include integrating information from Twitter on self-
language usage, avatar, emoticon, and responsive
patterns (Wei et al., 2017) or on profile attributes and
language (Sumner et al., 2012); integrating text,
image, and users' meta features from Twitter and
Instagram (Skowron et al., 2016); integrating
demographic features, social network activities, and
language extracted from Facebook, Twitter, and
YouTube (Farnadi et al., 2016); analysing text and
pictures contained in traces left in multiple public and
private social media platforms (Azucar et al., 2018).
Researchers agree that personality traits and
individual behaviour patterns are strongly related
(Augstein et al., 2019). In this direction, approaches
integrate information on networks-related behaviour
and personal traits (Lee & Kim, 2017; Zhao & Zhu,
2019); personal traits and emotions as a behaviour
regulating factor (Lerner et al., 2015; Tkalčič, 2020);
personal traits and behaviour activities of smart-
phone users (Stachl, Au, et al., 2020), and others.
Various personalized DSSs and recommendation
systems exploit digital traces as a resource of
personality information (e.g., (Courtin & Tomasena,
2016; Narducci et al., 2019)). Comprehensive
reviews of personality-based recommendation
systems that use information on personality traits
extracted from digital traces and combine it with user
behaviour can be found in (Augstein et al., 2019;
Suhaim & Berri, 2021).
Multiple personalized and adaptive systems rely
upon user types (or stereotypes) characterized with
common personal features. Researches that support
prediction of user types through classification include
a classification of social media users into types that
reflect the level of their engagement in the media
usage (Lee & Kim, 2017), personality classification
based on Twitter text (Pratama & Sarno, 2015), and
others. Ontologies, which provide classification
service, support a user categorization based on
personality traits, facets, culture, and age in relation
to specific tasks (El Bolock et al., 2020); personality
traits and facets recognized in different tests (Garcia-
Velez et al., 2018); textual data contained in the
digital traces (Alamsyah et al., 2021); combinations
of text and social behavioural aspects of user on
multiple social media (Sewwandi et al., 2017), and
many others.
The present research relies upon a user digital life
model as a collection of user-specific information.
This model is a structured representation of a part of
the content of user digital traces. Personality is
recognized through ontology-based classification. A
user ontology represents a priori knowledge on user
types. It infers a user type based on the type
definitions and information from the user digital life
model. The information used concerns the user traits
and online behaviour that this user manifest while
decision-making.
3 CONCEPTUAL FRAMEWORK
The conceptual framework of personalized decision
support based on user digital life model (Figure 1) is
intended to recommend decisions that the user would
made in the current situation (context). The main
components of this framework are user digital traces,
user profile, user digital life model, user segment,
user ontology, and context (Smirnov & Levashova,
2020). The user digital life model is the main source
of information for the rest of the components.
User profile is a set user characteristics that can
be used to create a descriptive portrait of an individual
and to identify one. User digital traces is a set of
records fixing information on the user activities
including decision-making. User digital life model is
a structured representation of a part of the content of
user digital traces, which carries information related
to the decision-making process of the user. User
segment is a group of users with common needs and
behavioural reactions when making decisions. User
ontology is a user model, which formalizes
knowledge to classify a user into a user type, i.e. into
a category of users distinguished by common
personality traits of these users as decision makers
and common decision-making behaviour. Context is
any information that characterizes the situation of the
user in the decision-making process. In the
conceptual framework, context comprises the user
identifying information and the information on the
user preferences, the user type, the problem requiring
CHIRA 2022 - 6th International Conference on Computer-Human Interaction Research and Applications
130
Figure 1: Conceptual framework for personalized decision support based on user digital life model.
a decision, and the knowledge domain that this
problem deals with.
When the user requests the DSS with a problem
requiring a decision, or when the DSS finds out that
the user needs a recommendation, the system infers
the user type and recommends a decision based on the
knowledge about the kinds of decisions customary for
the users of this type.
The framework components are modelled using
the set-theoretic approach.
User profile (UP):
=
(
_,_,_()
)
,
_()= _()
⋃
()
⋃
(),
where _ is the unique user identifier, _ is
the set of context-independent user characteristics;
_() is the set of context-sensitive user
characteristics in the context (); _
(
)
is
the user type in the context (); () is the set of
user preferences in the context ();
() is the set
of context-sensitive user characteristics other than the
user type and the user preferences (e.g., the user
location, local time, etc.); T is the period of existence
of the context C.
User digital life model (DL):
=(_,
(
,
)
,
(
,
)
,
{
(
,
)}
,
(
)
,
,
,
),
∈×,
∈×
,
∈
(
,
)
×,
where is the kind of the decision-making
problem (
(
,
)
means that the user
addresses the problem in the interval
(
,
)
);
is the time instant when the user starts
decision-making;
is the time instant when the user
has made a decision; is the knowledge
domain that the problem deals with
(
(
,
)
means that the domain knowledge
is dealt with in the interval
(
,
)
);
(
,
)
is the action carried out in the interval
(
,
)
(
≤
,
<
,
<
)
;
(
)
is the decision
made at the time instant
.
User segment (S):
=
(
,_,,
,
)
,
∈× ,
∈×
_,
where _ is the type of users sharing a
common behavioural pattern when choosing a
decision, is the set of behavioural variables
providing data to the behavioural pattern.
User ontology (
):
=
(
,,
)
,
=
⋃
, =
⋃
_
,
where is the set of ontology classes, is the set
of class relationships (→× ), is the
class that represents the user types,
=\,
is the set of ontology axioms,
_
is the set of
axioms that define the membership of the class
by a user,
=\
_
.
Context ():
(
)
=(_,_(),(),
(
)
,
(),
), (1)
_⊂_ , _()=
(
,
)
, ⊂_ , ⊂
_ , ()⊂ ,
()⊂ ,
()⊆ ,
∈
() × ,
where _ is the unique user identifier;
_() is the user type in the context
(
)
;
(
)
is the problem for that the user is
making a decision in the context
(
)
;
(
)
is the knowledge domain that the problem
(
)
deals with;
() is the set of user
preferences in the context
(
)
; =
(
,
)
.
User ontology
Contex
t
User profile
User digital life model
User segment
Type
User
Digital life
Digital traces
Axioms
define
has
describe
is of
p
rovides data
Problem
addresses
p
rovide data
p
rovide data
Preferences
has
includes
provides data
p
rovides data
Domain
deals with
provides data
Context-aware Personalized Decision Support based on User Digital Life Model
131
4 USER ONTOLOGY
The user ontology represents a priori knowledge on
user types and supports the user classification.
4.1 User Types
A user type is compound. It combines the type of
users as decision makers and the type of users as
online consumers. Two direct subclasses of the class
Type, that are DM_Type and Consumer_Type,
represent these subtypes, respectively. The user types
are context-sensitive.
The class DM_Type represents types of the users
depending on their decision-making styles and class
axioms that specify user characteristics influencing
these styles. A decision-making style affects
personality traits of a decision maker (El Othman et
al., 2020), preferences for the selection of an
alternative (Sharma & Pillai, 1996), and eventually
the decision. Kinds of the decision-making styles are
adopted from the management domain. According to
them, decision makers can be spontaneous, rational,
inert, risky, and cautious.
Detailed descriptions of the decision-making
styles (Allen, 2017; Bavol’ár & Orosová, 2015;
Sharma & Pillai, 1996) allowed us to find out factors
that influence decision making and that can be scored
based on information from digital traces (Table 1).
Number of decision makers expresses a preference
for an individual or collective decision-making.
Decision making time expresses the thoroughness
of the analysis and evaluation of the alternatives (this
time includes the time of searching for information).
It is scored as low, medium, or high.
Confidence degree scores the confidence of the
user in his/her knowledge and assessments in the
scale of low, medium, and high.
Complexity of decision-making procedure is the
complexity of the process thought that the user has to
go to reach the final decision (searching for
information, analysing and evaluating the
alternatives, consulting, and decision coordinating).
This procedure is proposed to be assessed as simple,
medium, or complex.
Criterion is the preference criterion (latent or
explicit) that the user applies to evaluate the
alternatives.
The class Consumer_Type classifies the users as
consumers of the recommended decisions based on the
behavioural variables (Var). The classification used
has come from the customer behaviour segmentation,
which focuses on the division of the users into groups
based on common online behaviour patterns. In the
paper, it is supposed that the users play the role of
Internet service consumers, the DSS is an online
service that recommends decisions, and the users’
behavioural patterns when they are thinking of either
accept a decision or do not reflect the online
behavioural patterns of these users as consumers.
Progressives, consolidators, always-hurrying,
traditionalists, and security-concerned consumers are
represented in the class Consumer_Type.
Based on detailed descriptions of the consumer
types, behavioural variables are identified, values for
that can be found in digital traces (
Table
2). The meanings of these variables are
intuitively clear; therefore, they are not described
unlike the factors.
4.2 User Digital Life Model
Actions and decisions specified in the model of user’s
digital life are analysed to identify the information
that can be used to score the factors (Table 3) and
variable values (Table 4).
Actions on information search (information
search requests) and communicating actions provide
information to score the number of decision makers
(f1) and to score the user confidence degree (f3) (a
decision maker that relies only on own knowledge, is
characterized by an extreme (high) degree of
confidence and opposite, a decision maker that looks
through large volumes of irrelevant information and
Table 1: Factors influencing decision making allocated to decision maker types.
Factor
Decision Maker Type
Spontaneous Rational Inert Risky Cautious
f1. Number of decision makers one-two group one-two one/group group
f2. Decision-making time low medium high medium high
f3. Confidence degree high medium low high low
f4. Complexity of decision-
making procedure
simple medium complex medium complex
f5. Criterion maximizing
rapidity of getting
benefit
maximizing
effectiveness of
problem resolving
maximizing
effectiveness of
problem resolving
maximizing
benefits
minimizing
losses
CHIRA 2022 - 6th International Conference on Computer-Human Interaction Research and Applications
132
Table 2: Behavioural variables influencing consumer type.
Variable
Consumer Type
Progressives Consolidators Always-hurrying Traditionalists Security-concerned
v1. Time spent in the Internet much moderate moderate little little
v2. Degree of involvement in
social networks
medium high do not use, usually low low
v3. Degree of Internet
services consuming
high high medium low low
v4. Degree of interest to
innovations
high medium low low low
v5. Loyalty level low medium high high medium
v6. Preferable
communication means
no
preferences
no
preferences
written voice voice
v7. Criterion maximizing
own benefits
maximizing
benefits of
other users
maximizing own
benefits,
minimizing time
utility
maximization
minimizing losses
contacts other individuals for help is characterized by
a low degree of confidence). Time values fixed for the
actions on information search and on the interactions
of the user with the DSS are used to calculate how
much time the user gathers and analyses information,
and evaluates the recommendation in order to either
accept or decline it (f2). All kinds of actions used to
score factors f1–f3 are used to score complexity of the
decision-making procedure (f4) (if the decision is
made quickly (the decision-making time is low) then
the procedure is assessed as simple; if the decision-
making time is high then the procedure is evaluated
as complex). The result of an analysis of the user
decisions is the criterion (f5), which expresses
explicit or latent preferences of the user.
Table 3: Digital life model as source for factor scores.
Factor Information from digital life model
f1 Number of individual recipients or groups to
which the user sent requests
f2 Time taken to analyze the request results, and
to evaluate the recommendation
f3 Kinds, number, and relevance of the
knowledge sources used
f4 All the above
f5 Decision
Table 4 provides information from the user digital
model that can be analysed to instantiate the
behaviour variables. Durations calculated based on
the time values fixed for the corresponding actions
instantiate the time behaviour variables (the first three
rows in the table). The degree of the Internet-services
consuming is calculated as the ratio of the number of
the kinds of the services used to the number of hits
normalized to low, medium, or high. The loyalty level
is determined by the information showing the user
interest to similar services offered by different
organizations. Communication activities provide
information about phone calls or written messages of a
user, on the basis of which the frequency of both is
calculated and the preferred means of communication
is determined. The criterion is determined based on the
information about the decisions that the user digital life
model specifies.
Table 4: Digital life model as source for variable values.
Variable Information from digital life model
v1 Time fixed for the on-line activities
v2 Time fixed for the social networks activities
v4 Time fixed for service hints comparing with
time of the service releases
v3 Kinds of services used, service hits
v5
• Activities on searching for the services that
are the same as offered by a specific site
• Activities on the usage of services that are
the same as offered by a specific site services
on other sites
• Activities on the usage of services offered
by competitors
v6 Communication activities
v7 Decision
5 PERSONALIZED DECISION
SUPPORT SCENARIO
The scenario of personalized decision support based
on user digital life model implements the general idea
of the conceptual framework and leaves aside cases
when the user does not accept the recommendation.
Such cases create a reason to a refinement of user
types and system learning, but they are out of research
Context-aware Personalized Decision Support based on User Digital Life Model
133
scope so far. The scenario considers situations that the
DSS addresses on the interval from the time instant
when the user starts decision-making until a
recommendation has been provided (Figure 2).
A problem that the user faces (the decision-making
problem) initiates the scenario. The information on
the unique user identifier _, the time t when
the user performs the actions causing the system
reaction, these actions
(
,
)
,
(
,
)
⊆
(
,
)
(the very first
action in time is
(
,
)
), the decision-making
problem problem(t), and the domain () is
identified in the user digital traces and becomes
represented in the user digital life model.
A formal problem model is built. The DSS uses
this model to solve the problem
(
)
as a
decision support problem. The function M:
problem(t) PM assigns the problem its formal
model.
Scores for the factors determining decision-
making styles and values for the behaviour variables
are assessed based on the user characteristics
Figure 2: Scenario of personalized decision support.
contained in the user profile and the information from
the user digital life model. These scores and values
fully or partly instantiate the ontology axioms and
rules that define the user types. The fully instantiated
axioms and rules become assertions (axioms and rules
describing individuals (Glimm et al., 2012)).
Based on the set of the assertions, the ontology
solves the classification problem and derives the user
type (user_type(t)) made up of the user type as
decision maker and the user type as consumer.
The information on the user identifier (user_ID),
the user type (user_type(t)), the kind of the problem
that the user addresses (problem(t)), the knowledge
domain (domain(t)), and the set of user preferences
(Pr
u
(t)) instantiates the context model C(T) (1).
Using the model (PM), the DSS solves the
problem problem(t) as a decision support problem.
The result of problem solving is a set of alternatives.
In accordance with the user type as consumer, the
segment of the users of this type is distinguished and
the kind of decisions customary to the users of this
segment is identified. Based on this kind of decisions
a recommended decision from the set of alternatives
is selected. The recommended decision is delivered to
the user, appears in the user digital traces, and saved
in the user digital life model:
(
)
=(_,
(
)
,
(
)
,
{
(
)}
,
(
)
).
6 USE CASE
The scenario above is demonstrated by an example of
decision support for the user that searches a book on
programming in Java in a library.
The user name is Alex; the part of the digital life
model built for Alex based on his digital traces for the
problem in question is as follows:
(
,
)
=
=
(
,ℎ
(
)
,
(
)
,
{
(
,
)})
=2020–11–19 19:55:16.057 , {
(
,
)
}=
=
. . ℎ(2020-11-19 19:55: 17.0648,2020-11-19 19: 55:17.926)
ℎ. . (2020-22-19 19: 55:16.057, 2020-11-19 19:55: 18.873)
. .Programming in Java(2020-11-19 19: 55:19.203, 2020-11-19 19:55: 19.936)
.
The model above specifies that the user identified
as Alex (_ = Alex) addresses the problem of
searching (
(
)
= ℎ
(
)
) a book
entitled “Programming in Java” in the library
(
(
)
=
(
)
).
The preferences of Alex coming from his profile
for the problem of book searching declare that the
preferable language for Alex is English. Based on the
scores for the factors influencing decision-making
styles and values for the behaviour variables, the user
Decision-making
problem arises
Collecting user-
specific information
Component of
conceptual framework
User digital
life model
Problem
formalization
User
profile
User type
recognition
Context model
instantiation
Problem solving
User segment
identification
Providing
recommendation
p
roblem(t)
M
: problem(t) PM
user_ID, Pr
u
(t), factor
scores, variable values
p
roblem(t),
domain (t)
C(T)
a set of alternatives
customary decisions
actions,
decisions
action(t
-
, t
+
)
Situation
User
recommended
decision
information flow
recommended
decision
User
segment
user_type(t)
P
_in(t),
P
_out,
Pr
u
(t)
user_ID, Pr
u
(t),
user_type(t),
CHIRA 2022 - 6th International Conference on Computer-Human Interaction Research and Applications
134
ontology classifies Alex as a spontaneous decision
maker and as an always-hurrying consumer. The
scores are obtained as results of an analysis of the
actions and decisions represented in Alex’s digital life
model. All the actions and decisions that concern the
problem of books searching are analysed, not just
those for the given time (for the time interval [t
0
, t
+
]
this model does not contain a decision).
Below, the instantiated context model (1) is given.
(
)
=
(,(
),(
),ℎ
(
)
,:English
(
)
),
SA = (spontaneous, always-hurrying),
=[
,
),
– the moment of the context model instantiation.
A set of available in the library books in English
devoted to programming in Java is the result of
problem solving. The DSS uses the information that
the always-hurrying users prefer quickly
implementable decisions to select a recommendation
from this set.
Among various options that the library suggests
are benefits for its subscribers. The DSS checks in the
Alex’s profile that he is a library subscriber and
recommends the book: “Herbert Schildt, Java: The
Complete Reference, Ninth Edition, McGraw-Hill
Education – Europe, 2014, 1312 p. (English,
Paperback), ISBN: 9780071808552.” This book is
digitalized and the library can provide access to a
digital copy of the book immediately after the request
of a subscriber. Alex accepts the recommendation by
ordering the book from the library.
Alex’s digital life model is updated with the
recommended decision:
=(,ℎ
(
,
)
,
(
,
)
,
{
(
,
)}
,"9780071808552" ),
(
(
)
=9780071808552).
7 CONCLUSIONS
Personalized decision support needs user-specific
information. Digital traces is a valuable resource of
such information. The paper proposes a conceptual
framework of personalized decision support in which
the core source of information about user specifics is
user digital life model. This model systematizes and
contextualizes the content of digital traces that can be
used to recognize personal user characteristics. Based
on these characteristics the user ontology infers a user
type. The DSS implementing the proposed
framework refers to the user type to recommend a
decision customary for the users of this type.
The novelties of the presented research are that it
proposes a user digital life model, which offers a new
means for organizing weakly structured content of
digital traces; suggests a way to recognise context-
sensitive user types from their digital traces, and gives
idea of a system dealing with context-sensitive user
types to provide personalized recommendation.
Future research can address the development of
decision support scenarios for the cases when the user
does not accept the recommendation.
ACKNOWLEDGEMENTS
The analysis of related research and the conceptual
framework for personalized decision support based
on user digital life model (Sections 2, 3) are due to the
grant from RFBR no. 20-07-00455, the user ontology
(Section 4) and the scenario implementing the
framework (Section 5) are due to the grant from
RFBR no. 20-07-00490, the user digital life model
derived from digital traces (Section 3) is due to State
Research (project no. FFZF-2022-0005).
REFERENCES
Alamsyah, A., Dudija, N. & Widiyanesti, S. (2021). New
Approach of Measuring Human Personality Traits
Using Ontology-Based Model from Social Media Data.
Information, 12(10), 413. https://doi.org/10.3390/
info12100413.
Allen, M. (Ed.). (2017). Spontaneous Decision Making. In
The SAGE Encyclopedia of Communication Research
Methods. SAGE Publications, Inc. https://doi.org/10.41
35/9781483381411.n583.
Augstein, M., Herder, E. & Wörndl, W. (2019).
Personalized Human-Computer Interaction. De
Gruyter. https://doi.org/10.1515/9783110552485.
Azucar, D., Marengo, D. & Settanni, M. (2018). Predicting
the Big 5 personality traits from digital footprints on
social media: A meta-analysis. Personality and
Individual Differences, 124, 150–159. https://doi.org/
10.1016/j.paid.2017.12.018.
Bavol’ár, J. & Orosová, O. (2015). Decision-making styles
and their associations with decision-making
competencies and mental health. Judgment and
Decision Making, 10(1), 115–122. http://journal.sjdm.
org/14/141103/jdm141103.pdf.
Breiter, A. & Hepp, A. (2018). The Complexity of
Datafication: Putting Digital Traces in Context. In
Communicative Figurations (pp. 387–405). Springer
International Publishing. https://doi.org/10.1007/978-
3-319-65584-0_16.
Courtin, C. & Tomasena, M. (2016). A benchmarking
platform for analyzing corpora of traces. Proceedings
of the 3rd IEEE/ACM International Conference on Big
Data Computing, Applications and Technologies, 178–
186. https://doi.org/10.1145/3006299.3006303.
Context-aware Personalized Decision Support based on User Digital Life Model
135
El Bolock, A., Herbert, C. & Abdennadher, S. (2020).
CCOnto: Towards an Ontology-Based Model for
Character Computing. In F. Dalpiaz, J. Zdravkovic &
P. Loucopoulos (Eds.), Research Challenges in
Information Science. RCIS 2020 (Vol. 385, pp. 529–
535). Springer. https://doi.org/10.1007/978-3-030-
50316-1_34.
El Othman, R., El Othman, R., Hallit, R., Obeid, S. & Hallit,
S. (2020). Personality traits, emotional intelligence and
decision-making styles in Lebanese universities
medical students. BMC Psychology, 8(1), 46.
https://doi.org/10.1186/s40359-020-00406-4.
Farnadi, G., Sitaraman, G., Sushmita, S., Celli, F.,
Kosinski, M., Stillwell, D., Davalos, S., Moens, M.-F.
& De Cock, M. (2016). Computational personality
recognition in social media. User Modeling and User-
Adapted Interaction, 26(2–3), 109–142.
https://doi.org/10.1007/s11257-016-9171-0.
Garcia-Velez, R., Galan-Mena, J., Lopez-Nores, M. &
Robles-Bykbaev, V. (2018). Creating an Ontological
Networks to Support the Inference of Personality Traits
and Facets. 2018 IEEE XXV International Conference
on Electronics, Electrical Engineering and Computing
(INTERCON), 1–4. https://doi.org/10.1109/INTER
CON.2018.8526453.
Glimm, B., Horrocks, I., Motik, B., Shearer, R. & Stoilos,
G. (2012). A novel approach to ontology classification.
Journal of Web Semantics, 14, 84–101.
https://doi.org/10.1016/j.websem.2011.12.007.
Lee, J. & Kim, G. (2017). Social media user classification.
Proceedings of the International Conference on
Electronic Commerce, 1–3. https://doi.org/10.1145/31
54943.3154963.
Lerner, J. S., Li, Y., Valdesolo, P. & Kassam, K. S. (2015).
Emotion and Decision Making. Annual Review of
Psychology, 66(1), 799–823. https://doi.org/10.1146/
annurev-psych-010213-115043
Narducci, F., Basile, P., Iovine, A., De Gemmis, M., Lops,
P. & Semeraro, G. (2019). A Framework for building
Chat-based Recommender Systems. In A. Cuzzocrea,
F. Bonchi & D. Gunopulos (Eds.), Proceedings of the
CIKM 2018 Workshops co-located with 27th ACM
International Conference on Information and
Knowledge Management (CIKM 2018) (Vol. 2482).
http://ceur-ws.org/Vol-2482/paper47.pdf.
Pratama, B. Y. & Sarno, R. (2015). Personality
classification based on Twitter text using Naive Bayes,
KNN and SVM. 2015 International Conference on
Data and Software Engineering (ICoDSE), 170–174.
https://doi.org/10.1109/ICODSE.2015.7436992.
Rentfrow, P. J. & Gosling, S. D. (2003). The do re mi’s of
everyday life: The structure and personality correlates
of music preferences. Journal of Personality and Social
Psychology, 84(6), 1236–1256. https://doi.org/10.1037/
0022-3514.84.6.1236.
Sewwandi, D., Perera, K., Sandaruwan, S., Lakchani, O.,
Nugaliyadde, A. & Thelijjagoda, S. (2017). Linguistic
features based personality recognition using social
media data. 2017 6th National Conference on
Technology and Management (NCTM)
, 63–68.
https://doi.org/10.1109/NCTM.2017.7872829.
Sharma, A. & Pillai, R. (1996). Customers’ Decision-
Making Styles and their Preference for Sales Strategies:
Conceptual Examination and an Empirical Study.
Journal of Personal Selling & Sales Management,
16(1), 21–33. https://doi.org/http://dx.doi.org/10.1080/
08853134.1996.10754042.
Skowron, M., Tkalčič, M., Ferwerda, B. & Schedl, M.
(2016). Fusing Social Media Cues. Proceedings of the
25th International Conference Companion on World
Wide Web - WWW ’16 Companion, 107–108.
https://doi.org/10.1145/2872518.2889368.
Smirnov, A. & Levashova, T. (2020). Intelligent Decision
Support Based on User Digital Life Model: Principles
and Conceptual Framework. 2020 27th Conference of
Open Innovations Association (FRUCT), 247–252.
https://doi.org/10.23919/FRUCT49677.2020.9211055.
Stachl, C., Au, Q., Schoedel, R., Gosling, S. D., Harari, G.
M., Buschek, D., Völkel, S. T., Schuwerk, T.,
Oldemeier, M., Ullmann, T., Hussmann, H., Bischl, B.
& Bühner, M. (2020). Predicting personality from
patterns of behavior collected with smartphones.
Proceedings of the National Academy of Sciences,
117(30), 17680–17687. https://doi.org/10.1073/
pnas.1920484117.
Stachl, C., Pargent, F., Hilbert, S., Harari, G. M., Schoedel,
R., Vaid, S., Gosling, S. D. & Bühner, M. (2020).
Personality Research and Assessment in the Era of
Machine Learning. European Journal of Personality,
34(5), 613–631. https://doi.org/10.1002/per.2257.
Suhaim, A. Bin & Berri, J. (2021). Context-Aware
Recommender Systems for Social Networks: Review,
Challenges and Opportunities. IEEE Access, 9, 57440–
57463. https://doi.org/10.1109/ACCESS.2021.3072165.
Sumner, C., Byers, A., Boochever, R. & Park, G. J. (2012).
Predicting Dark Triad Personality Traits from Twitter
Usage and a Linguistic Analysis of Tweets. 2012 11th
International Conference on Machine Learning and
Applications, 386–393. https://doi.org/10.1109/ICM
LA.2012.218.
Tkalčič, M. (2020). Complementing Behavioural Modeling
with Cognitive Modeling for Better Recommendations.
In D. Helic, G. Leitner, M. Stettinger, A. Felfernig & Z.
W. Raś (Eds.), Foundations of Intelligent Systems.
ISMIS 2020 (Vol. 12117, pp. 3–8). Springer.
https://doi.org/10.1007/978-3-030-59491-6_1.
Wei, H., Zhang, F., Yuan, N. J., Cao, C., Fu, H., Xie, X.,
Rui, Y. & Ma, W.-Y. (2017). Beyond the words:
predicting user personality from heterogeneous
information. Proceedings of the Tenth ACM
International Conference on Web Search and Data
Mining, 305–314. https://doi.org/10.1145/3018661.30
18717
Zhao, N. & Zhu, T. (2019). Psychological .Profiles
Prediction Using Online Social Network Behavior
Data. In Z. Yan (Ed.), Analyzing Human Behavior in
Cyberspace (pp. 75–87). IGI Global. https://doi.org/
10.4018/978-1-5225-7128-5.ch005.
CHIRA 2022 - 6th International Conference on Computer-Human Interaction Research and Applications
136
