Prediction of Human Personality Traits From Annotation Activities

Nizar Omheni

, Omar Mazhoud

, Anis Kalboussi

and Ahmed HadjKacem

Higher Institute of Computer and Management, University of Kairouan, ReDCAD Laboratory,

Khmais Alouini Street 3100, Kairouan, Tunisia

Faculty of Economics and Management, University of Sfax, ReDCAD Laboratory,

Road of the Airport Km4 3018, Sfax, Tunisia

Keywords:

Annotation, Personality, Big Five Personality Model.

Abstract:

We show how reader’s annotation activity captured during an active reading session relates to their personal-

ity, as measured by the standard Five Factor Model. For 120 volunteers having usually the habit of reading,

we gather personality data and annotation practices. We examine correlations between readers personality

and such features of their annotative activities such as the total number of annotation acts, average number of

annotation acts, number of textual annotation acts, number of graphical annotation acts, number of referential

annotation acts and number of compounding annotation acts. Our results show signiﬁcant relationships be-

tween personality traits and such features of annotation practices. Then we show how multivariate regression

allows prediction of the readers personalities traits given their annotation activities.

1 INTRODUCTION

Studying human activity and interaction with technol-

ogy has grown dramatically over the last decade. Yet

studying reading poses particular challenges. (Mar-

shall,2010) reported the citation of Tzvetan Todorov,

quoted by Nicholas Howe in Jonathan Boyarins com-

pilation, the Ethnography of Reading: ”Nothing is

more commonplace than the reading experience, and

yet nothing is more unknown. Reading is such a mat-

ter of course that at ﬁrst glance it seems there is noth-

ing to say about it”. Although details of reading ac-

tivity (moving eyes, writing annotation...) may tell us

something about the reader. When people read and

interact actively with their reading materials they do

unselfconscious activities which can be keys features

to their personalities.

For decades, the psychologists search to under-

stand the human personality and to ﬁnd a systematic

way to measure it. After several researches they show

a relation of dependence between human personality

traits and different behaviors. (Ryckman,2010) re-

ported the Allport’s

deﬁnition of personality: ”per-

sonality is the dynamic organization within the in-

Gordon Willard Allport (November 11, 1897 October

9, 1967) was an American psychologist. He was one of the

ﬁrst psychologists to focus on the study of the personality,

and is often referred to as one of the founding ﬁgures of

personality psychology.

dividual of those psychophysical systems that deter-

mine his characteristic behavior and thought”. Thus,

in Allports view human behavior is really controlled

by internal forces known as the personality traits.

This paper attemps to bridge the gap between

reading activity research and personality research

through reader’s annotation practices. Our core re-

search question asks whether annotation activity can

predict personality traits. If so, then there is an oppor-

tunity to use a natural human practice as a new source

to better understand the reader personality.

Several works has shown the opportunity of pre-

dicting user personality using the information peo-

ple reveal in their online social proﬁle (Twitter, Face-

Book) (Bachrach et al,2012) (Golbeck et al,2011).

They refer to what people share, self-description, sta-

tus updates, photos, tags, etc. We pretend that annota-

tive activity is more spontaneously and natural prac-

tice and it can reveal something about human person-

ality.

Personalization attracted increased attention in

many areas. So the need to predict personality traits

increases over time mainly when several research has

shown the link between personality traits and suc-

cess in human relationship and practices (Barrick and

Mount,1991) (Eswaran et al,2011). By nature an in-

troverted person is not interested to make so much re-

lation with other while an extravert person do. Actu-

ally, certain developed recommendation systems con-

263

Omheni ., Mazhoud O., Kalboussi A. and HadjKacem A..

Prediction of Human Personality Traits From Annotation Activities.

DOI: 10.5220/0004801302630269

In Proceedings of the 10th International Conference on Web Information Systems and Technologies (WEBIST-2014), pages 263-269

ISBN: 978-989-758-024-6

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

sider the personality traits as key feature of recom-

mendation (Nunes et al,2008).

The paper is structured as follows: In Section 2,

we present background on the Big Five personality

index. Then we present our experimental setup and

methods. In the third section we present the results on

correlation between each annotative activity feature

and personality factor. Next, we show how multivari-

ate regression allows prediction of annotators person-

alities traits given their annotation activities. We con-

clude with a discussion of the possible implications

that this work has for such domains of application.

2 BACKGROUND AND RELATED

WORK

2.1 The Big Five Personality Model

The big ﬁve personality traits are the best ac-

cepted and most commonly used scientiﬁc measure

of personality and have been extensively researched

(Peabody and De Raad,2002). That personality is

well described as ﬁve traits was discovered through

the study of the adjectives from natural langage that

people used to describe themselves and then analyz-

ing the data with a statistical procedure known as fac-

tor analysis that is used to reduce lots of information

down to its most important parts. In the following we

cite a brief explanation of the ﬁve personality traits.

2.1.1 Openness to Experience

Openness includes traits like imagination, apprecia-

tion for art, depth of emotions, adventure, unusual

ideas, intellectual curiosity, and willingness to experi-

ment. People who score high in openness like usually

to learn new things and enjoy new experiences.

2.1.2 Conscientiousness

Conscientiousness includes traits like orderliness,

selfdiscipline, deliberateness, and striving the

achievement. People that have a high degree of

conscientiousness are planned, have the tendency

to act dutifully, have the sense of responsibility and

competence.

2.1.3 Extraversion

Extraversion includes traits like energy, positive

emotions, surgency, assertiveness, sociability and

talkativeness. Extraverts people get their energy from

interacting with others, while introverts get their en-

ergy from within themselves.

2.1.4 Agreeableness

Agreeableness includes traits like trust in others, sin-

cerity, altruism, compliance, modesty and sympathy.

People that have high degree of agreeableness are

friendly, cooperative, and compassionate, while peo-

ple with low agreeableness may be more distant.

2.1.5 Neuroticism

Neuroticism relates to ones emotional stability and

degree of negative emotions. This dimension mea-

sures the people degree of anxiety, angry, moodiness,

and the sensitivity to stress. People that score high

on neuroticism often experience emotional instability

and negative emotions.

2.2 Related Work

In (Burger,2011) view the personality is a ”consistent

behavior patterns and intrapersonal processes orig-

inating within the individual”. Trait psychologists

assume that personality is relatively stable and pre-

dictable (Burger,2011). So, several research work has

been done with personality traits as it inﬂuences hu-

man decision making process and interests. (Nunes

et al,2008) pionneered the model and implement of

personality traits in computers. Indeed, (Nunes et

al,2008) propose to model the user’s traits in a pro-

ﬁle which they called User Psychological Proﬁle -

UPP. In order to ﬁll in the proﬁle UPP the authors

utilised an online tool called the NEO-IPIP

inven-

tory based on 300 items. Through user’s answers to

NEO-IPIP inventory the authors are able to predict

the user personality. Through their experimentation

(Nunes et al,2008) try to prove that Recommender

Systems can be more efﬁcient if they use the User

Psychological Proﬁle (UPP). Although the authors

follow an explicit way to predict the user traits, the re-

sults presented in (Nunes,2008) are fruitful. (Tkalcic

et al,2009) propose a personality-based approach that

is based on the big ﬁve model for collaborative ﬁl-

tering Recommender Systems. In fact, the authors

calculate the user personality scores by means of a

questionnaire. Then they measure the user similar-

ity, that is based on personality, that yields a list of

close neighbours. This list is used after as a database

to compile a personalized list of recommended items.

The NEO-IPIP is a computer based Personality Inven-

tory, able to measure people Personality Traits created by

John Johnson (Johnson.).

WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies

264

(Roshchina et al,2011) propose personality-based rec-

ommender system which the aim is to select for the

user reviews that have been written by like-minded in-

dividuals. The user similarity is calculated based on

the presonality traits according to the Big Five model.

The authors predict the presonality traits based on lin-

guistic cues collected from the user-generated text.

(Bachrach et al,2012) and (Golbeck et al,2011) show

the relationship between personality traits and various

features of social media (FaceBook, Twitter). Their

ﬁndings prove the possibility to predict accurately the

user’s personality through the publicly avaible infor-

mation on their social network proﬁle.

As it is mentioned above most of works has been

done with the ”Big Five” model of personality dimen-

sions which has emerged as the most well-researched

and regarded measures of personality in last decades.

The best of our knowledge, our work is among the

ﬁrst to look at the relationship between annotation ac-

tivity and personaltiy traits. Much works try to pre-

dict personality from what the user offer consciously

(answers to a questionnaire, informations available on

public social proﬁle...). Despite the fact that the ﬁnd-

ings of these researchs are fruitful we believe that pre-

dicting personality from annotative activity is more

credible as the annotation is deﬁned as ”a basic and

often unselfconscious way in which readers interact

with texts” (Marshall,2010).

Due to the spontaneously and unselfconscious as-

pects of annotation we are interested to predict per-

sonality from this potential source of knowledge.

3 DATA COLLECTION

We consider group of 120 volunteers. The subjects se-

lected were recruited with respect to certain criterias.

Infact, the age of our volunteers should be between 18

and more and they should be academic people. In our

sample we have the two sex (44 women and 76 men).

Another criteria for selection , we asked if the volun-

teer has the habit of reading and does he annotated his

documents frequently. If all these conditions exist the

subject can be selected to our experimentation.

Each subject was instructed to answer a standard Five

Factor Model questionnaire (the NEO-IPIP Inven-

tory). Then, he obtained a feedback regarding his per-

sonality based on his responses. This step gives us

the personality scores based on the Big Five Model

for each volunteer. To associate personality scores

to subjects annotative activities, we gather annotation

practices for each people. Here, we collect documents

annotated in a spontaneous and natural way. So we

asked, ﬁrst of all, if the subject had a document an-

notated previously (academic course, book,...). If not

we asked him what topics interest him, then we give

him an article with few pages to not weary him.

We are very careful to the comfortability of the vol-

unteers during the experience to guarantee their spon-

taneous and natural reactions. Thus they are free to

choose places and conditions to read and annotate the

documents and they have enough time to do. The

strategy followed give us fruitful results. Infact, the

different subjects (who have not a document anno-

tated previously) interact actively with the reading

materials in view of the feel of comfortableness and

the interest to the document read.

Figure 1: Annotation practices of a reader.

3.1 User Annotation Activity

(Marshall,2010) deﬁnes the annotation as ”a basic and

often unselfconscious way in which readers interact

with texts”. We mean by annotation the act to add

a critical or explanatory notes to a written work, to

highlight a passage, to write down, and so on marks

the reader makes on a page during his reading activity.

To fulﬁl our experimentation, we ask each subject to

give us an annotated paper document. Then, we anal-

yse the readers annotations to extract some features.

We started by classifying annotations in the three

general categories cited by (Agosti and Ferro,2003) :

graphical annotation acts, textual annotation acts, and

reference annotation act all depends to the material-

ization sign of the semantics of the annotation added

to the annotated document. Then, for each reader, we

collect a simple set of statistics about their annotative

activity. These included the following:

1. Total Number of Annotation Act (TNAA)

2. Average Number of Annotation Act (num-

ber of annotation acts per a single annotated

page)(ANAN)

3. Number of Graphical Annotation Act (NGAA)

4. Number of Textual Annotation Act (NTAA)

PredictionofHumanPersonalityTraitsFromAnnotationActivities

265

5. Number of Reference Annotation Act (NRAA)

6. Number of compounding Annotation Act (textual

sign, graphic sign and reference sign of annotation

act can be compounded together in order to ex-

press complex meanings of annotation)(NCAA).

This set of statistics tends to characterize quantita-

tively the reader’s annotation practices. Next we run

a Pearson correlation

analysis between subjects’ per-

sonality scores and each of the features obtained from

analyzing their annotative activities.

4 PERSONALITY AND

ANNOTATION FEATURES

CORRELATION

We study the Pearson correlation between subjects’

personality scores and each of the features obtained

from analyzing their annotative activities. We report

the correlation values in table I. Those that were sta-

tistically signiﬁcant for p < 0.05 are bolded.

Table 1: Pearson correlation values between annotation fea-

tures scores and presonality scores.

Open. Consc. Extra. Agree. Neuro.

TNAA -0,059 0,128 -0,138 0,089 -0,287

ANAA 0,003 0,080 -0,210 0,163 -0,183

NGAA -0,067 0,040 -0,130 0,105 -0,207

NTAA 0,001 0,182 0,040 0,085 -0,211

NRAA -0,075 0,045 -0,122 0,077 -0,207

NCAA -0,059 -0,012 -0,147 0,014 -0,219

We found in our analysis fewer signiﬁcant corre-

lations, but we believe a larger sample size would pro-

duce much better results. However, the results we ob-

tained even with a small sample show promise that the

annotative activity can be useful for computing such

personality traits. In fact, table I shows signiﬁcant

correlations for Neuroticism, Conscientiousness, and

Extraversion traits. We need larger sample to verify

the inference of the other traits from peoples annota-

tions.

Next, we present the scatter plots for the most sig-

niﬁcant correlations between annotation practices fea-

tures and personality traits. These plots presenting

the relationship between annotative activity features

and human traits, where horizontal axis represents the

average personality trait scores and the vertical axis

represnts the annotative activity feature values.

Pearson’s correlation r ∈ [−1,1] measures the linear re-

lationship between two random variables.

4.1 Conscientiousness

As presented in table I Conscientiousness is positively

related to the number of textual annotation act (ﬁg.2).

The rest of the correlation values are not considerated

because of p-value > 0.05. But this is not a reason to

reject deﬁnitively the rest of annotation features as a

larger sample size may produces other signiﬁcant cor-

relations.

The considered correlation may indicates that consci-

entious people are interested to use textual annota-

tion acts. Infact, conscientious individuals are prudent

which means both wise and cautious, better organized

and they avoid acting spontaneously and impulsively.

Thus, it may be the case that people who have high

degree of conscientiousness are interested to use tex-

tual annotation more than other annotation acts as it

demands more reﬂexion, reasoning and cognitive ef-

fort.

Figure 2: Scatter Plot showing Number of Textual Annota-

tion Act against Conscientiousness scores.

4.2 Extraversion

According to results shown in table I, Extraversion is

negatively correlated with the average number of an-

notation act (ﬁg.3). The rest of the correlation values

can be probably signiﬁcant with a larger sample size.

We can interpret the regression ﬁt shown in ﬁgure 3

as follow: The ﬁt is correlated negatively which is not

surprising as extraversion is marked by pronounced

engagement with the external world where extraverts

tend to be energetic and talkative while introverts are

more likely to be solitary and reserved. Thus, it may

be the case that reading and annotation is an intimate

activities, we do it in private, so people who are so-

cially active are less willing to practice annotation.

4.3 Neuroticism

Table I shows that neuroticism is negatively corre-

lated with all the features of annotation activity. Here,

WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies

266

Figure 3: Scatter Plot showing Average Number of Annota-

tion Act against Extraversion scores.

the chosen sample size is sufﬁcient to have signiﬁcant

correlations for all the annotation features. The differ-

ent correlation values are very signiﬁcant which can

show the sensitivity of annotation practices to the neu-

roticism trait.

In other hand, one possible explanation for these cor-

relations is that more Neurotic people are emotion-

ally reactive and they experience negative emotions

for unusually long periods of time which can dimin-

ish the neurotics ability to think clearly and make de-

cisions. Thus those who score high on Neuroticism

are less eager to use annotation act as they can not

actively and critically engaging with the content for a

long periods of time.

Figure 4: Scatter Plot showing Total Number of Annotation

Act against Neuroticism scores.

4.4 Openness and Agreeableness

Unfortunately, the correlation values related to the

Openness and Agreeableness traits are very low. But

we can not reject deﬁnetly the hypothese of prediction

of these traits from annotation activity. We may ob-

tain signiﬁcant values if we increase the sample size.

As an exemple, the p-value of the regression ﬁt of

the Average Number of Annotation Act against the

Agreeableness traits is p = 0.076. This value can be

ameliorated with a larger sample size.

Figure 5: Scatter Plot showing Number of compounding

Annotation Act against Neuroticism scores.

5 PREDICTING PERSONALITY

Previously we examined the correlations between

each of Big Five personality dimension and anno-

tative activity features. Now, we are interested to

make predictions about a subject’s personality based

on multiple annotation features. First of all, we used

the multivariate linear regression to predict each per-

sonality trait using the annotation features. Next, we

used the coefﬁcient of multiple determination R

measure the strength of ﬁt. Also, we measure the

F-test to verify the statistical signiﬁcance of the col-

lective inﬂuence that have the annotation features on

the personality traits. Thus, larger values of the F-test

statistic provide stronger evidence against H

. To re-

ject H

the value of the F-test should exceeds a critical

value calculated as follow:

F =

(1 − R

)/[n − (K + 1)]

Where k is the number of explanatory variables in our

model which corresponds to the number of annotation

activity features(K=6)and n represents the sample size

(n=120). So the F

observed

is compared against a critical

F with 6 degree of freedom in the numerator and n-7

degrees of freedom for error in denominator. In our

case F

critical

=2.18 for alpha level

of 0.05. Results

shown in table II indicate that the null hypothesis is

rejected for two cases. In fact, Neuroticism and Con-

scientiousness can be predicted with reasonable ac-

curacy using features of annotation activity, whereas

other traits are more difﬁcult to be predicted using an-

notations. Prediction regarding Conscientiousness is

reasonably accurate, with R

value of 0.12, F

observed

The null hypothesis states that there is no relationship

between annotation activity features and personality traits.

The alpha level is deﬁned as the probability of what is

called a Type I error in statistics. That is the probability of

rejecting H

when in fact it was true.

PredictionofHumanPersonalityTraitsFromAnnotationActivities

267

value of 2.52 which exceeds the F

critical

value and P-

value of 0.03 which is lower than the α value where

P-value is the probability the F-test statistic is larger

than the observed F-value. For Neuroticism we ob-

tained the model with the best ﬁt, with an R

value

of 0.14, F

observed

value of 3.11 and P-value of 0.01,

indicating quite accurate a prediction. The model for

Extraversion has a lower ﬁt and the model for Agree-

ableness is even less accurate. It seems that Openness

is the hardest trait to predict using annotation activity

features.

Table 2: Predicting personality traits using annotation ac-

tivity features through multivariate linear regression.

Trait R

F test P-value

Openness 0.03 0.57 0.76

Conscientiousness 0.12 2.52 0.03

Extraversion 0.07 1.32 0.25

Agreeableness 0.05 1.03 0.41

Neuroticism 0.14 3.11 0.01

6 DISCUSSION

In this study we show that Neuroticism and Conscien-

tiousness traits are correlated with annotation activity

features. We expect a larger sample size can be help-

ful to verify the correlation of the other human traits

to annotation practices.

Our ﬁndings are based on pen-and-paper approach

which is qualiﬁed by its relative ease with which the

reader may interact with a document in an intuitive

and familiar manner.

Recent researchs endeavor to replace the ”pen-

and paper” pardigm for the annotating needs. Dif-

ferent systems and tools of annotation are developed

such as: iAnnotate (Plimmer et al,2010), u-Annotate

(Chatti et al,2006), YAWAS

, iMarkup (2013), etc.

Such tools enable readers to annotate their digital doc-

uments with free form annotations similarly to ”pen-

and paper” case. iAnnotate for example is an anno-

tation tool for android system and it enables users to

add annotations with the pencil, highlighter, and note

tools.

Recently we intend talking about new products for

reading such as the tablet. With the aid of such de-

vices, the user may interact easily with a digital docu-

ment and enter her annotations as he do in the case of

”pen-and-paper”. Thus, our ﬁndings is promising and

original to be applied in such digital areas.

Yawas is a free web annotation tool for Firefox and

Chrome built on top of Google Bookmarks. Yawas enables

you to highlight text on any webpage and save it in your

Google Bookmarks account.

We believe it’s the occasion to develop a system

to attract those that have a curiosity to use anno-

tation platforms and practices - from end users in-

cluding scholars, scientists, journalists, public ser-

vants...etc. We expect our system enables readers to

interact actively with their reading materials via an-

notation practices. Our goal is to use the traced anno-

tations on the digital document to infer the annotator

personality traits. Thus the expected system should

contain free form annotation tool easy to be used, be

able to infer user personality traits from captured an-

notation activity and reﬁnes the user traits proﬁle by

reference to new captured annotations. Based on the

modelled proﬁle we expect our system be able to offer

services to users such as friendship recommendation

based on similarity of users personality, customizing

user interface using such predeﬁned personas, sharing

annotated documents...etc.

To achieve the personalization process we need to

know certain user’s features. Several research works

prove that prediction of personality traits reveals a

lot about a user’s features. These ﬁndings was ap-

plied in several domains such as recommender sys-

tems (Nunes et al,2008) (Roshchina et al,2011) and

the results is interesting.

The annotation as an unselfconscious practice

constitutes a credible source of knowledge. In

(Kalboussi et al,2013) the annotation activity is used

to invoke the appropriate Web services to users. This

proves that annotation is rich enough to be used dif-

ferently.

In this paper we use annotation to infer such

reader traits. That is a promising work and represents

a new tendency to model user personality from human

behaviour.

In other hand, let’s be objective, there are some

limitations to this work. The most important issue is

the sample size as we expect more signiﬁcant results

with a larger sample. This limitation may be due to

the dependency to ”pen-and-paper” approach which

prevents us to beneﬁt from the population of readers

over the web. In addition, people are not interested

to participate in our experimentation unless they are

motivated. We expect resolving these issues in future

works.

Finally, our work is the ﬁrst step to study the rela-

tion of reader annotation practices to human person-

ality traits. So much perspectives can be subjects of

future works such as studying factors which are likely

to inﬂuence annotating behaviour such as familiarity

with annotation tools and interest in the content topic.

Also our research can be extended to study the pos-

sibility to predict human traits through social annota-

tions. These avenues and others are very interesting

WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies

268

and represent an opened future directions which needs

more investigations.

7 CONCLUSION

In this paper, we have shown that such users’ per-

sonalities traits can be predicted from their annota-

tion practices. With this ability of prediction many

opportunities are opened which suggests future direc-

tions in variety of areas such as user modeling, recom-

mender systems, user interface design and so on areas

relative to personalization research domain. Further-

more, this work bridges the gap between the reading

and the personality research domains and it remains

an open research questions to see whether personality

can also be predicted using other potentially features

of reading activity as well as the inﬂuence of such en-

vironmental factors on human annotation behaviour.

REFERENCES

Agosti,M. Ferro,N., 2003. Annotations: Enriching a Digi-

tal Library. In: Proceedings of the 7th European Con-

ference (ECDL) Trondheim, Norway. Springer Pages

88-100.

Bachrach,Y. Kosinski,M. Graepel,T. Kohli,P. Still-

well,D.,2012. Personality and Patterns of Facebook

Usage. In: Web Science12 Proceedings of the 3rd

Annual ACM Web Science Conference. ACM New

York, NY, USA, Pages 24-32.

Burger, J. M., 2011. Personality. Editor Wadsworth, USA.

Barrick, M. R. Mount, M.K., 1991. The big ﬁve personal-

ity dimensions and job performance: a meta-analysis.

Personnel psychology, Volume 44, Issue 1, Pages 126.

Chatti, M. A. Sodhi,T. Specht, M. Klamma,R. Klemke,R.,

2006. u-Annotate: An Application for User Driven

Freeform Digital Ink Annotation of E-Learning Con-

tent. In: Proceedings of the Sixth IEEE Interna-

tional Conference on Advanced Learning Technolo-

gies (ICALT), Washington, DC, USA. Pages 1039-

1043.

Eswaran,S. Aminul Islam,Md. Dayang Hasliza,M.Y., 2011.

A Study of the Relationship between the Big Five Per-

sonality Dimensions and Job Involvement in a Foreign

Based Financial Institution in Penang. In: Interna-

tional Business Research. Vol. 4, No. 4.

Johnson, J.A. The IPIP-NEO: International

Personality Item Pool Representation

of the NEOPI-R, viewed 8 June 2013,

http://www.personal.psu.edu/ j5j/IPIP/ipipneo30

0.htm

Golbeck, J. Robles,C. Edmondson, M. Turner,K., 2011.

Predicting Personality from Twitter. In: Privacy, Se-

curity, Risk and Trust PASSAT, 2011 IEEE Third In-

ternational Conference on Privacy, Security, Risk, and

Trust, and 2011 IEEE Third International Confernece

on Social Computing.

Kalboussi, A. Mazhoud, O. Hadj Kacem,A.,2013. Annota-

tive Activity as a Potential Source of Web Service Invo-

cation. In: Proceedings of the 9th International Con-

ference on Web Information Systems and Technolo-

gies (WEBIST), Aachen, Germany. SciTePress Pages

288-292.

Nunes, M.A. Cerri,S.A. Blanc,N., 2008. Improving Rec-

ommendations by Using Personality Traits in User

Proﬁles. In: International Conferences on Knowledge

Management and New Media Technology, Graz, Aus-

tria.

Nunes, M.A. Cerri,S.A. Blanc,N., 2008. Towards User Psy-

chological Proﬁle. In: IHC ’08 Proceedings of the

VIII Brazilian Symposium on Human Factors in Com-

puting Systems. ACM Pages 196-203.

Nunes,M.A.,2008. Recommender Systems based on Person-

ality Traits. PhD Thesis, University of MONTPEL-

LIER 2.

Marshall, C., 2010. Reading and Writing the Electronic

Book.. Editor Gary Marchionini, University of North

Carolina, Chapel Hill.

Plimmer, B. Hsiao-Heng Chang, S. Doshi, M. Laycock,L.

Seneviratne,N., 2010. iAnnotate: Exploring Multi-

User Ink Annotation in Web Browsers. In: Proceed-

ings of the 11th Australasian User Interface Confer-

ence (AUIC). ACM Pages 52-60.

Peabody, D. De Raad, B., 2002. The Substantive Nature

of Psycholexical Personality Factors: A Comparison

Across Languages. Journal of Personality and Social

Psychology, vol 83, issu 10; Pages 983-997.

Roshchina, A. Cardiff, J. Rosso, P., 2011. User Proﬁle

Construction in the TWIN Personality-based Recom-

mender System.. In: Proceedings of the Workshop

on Sentiment Analysis where AI meets Psychology

(SAAIP), IJCNLP. Pages 7379.

Ryckman,R. M., 2008. Theories of Personality. Editor

Thomson Wadsworth USA.

Schmitt, D. Allik, J. McCrae, R. Benet-Martinez,V., 2007.

The geographic distribution of Big Five personality

traits: Patterns and proﬁles of human self-description

across 56 nations. Journal of Cross-Cultural Psychol-

ogy, vol 38, Pages 173-212.

Tkalcic, M. Kunaver, M. Tasic, J. Ko

sir,A., 2009. Per-

sonality based user similarity measure for a col-

laborative recommender system. In: Proceedings of

the 5th Workshop on Emotion in Human-Computer

Interaction-Real world challenges, Cambridge, UK.

Pages 30-37.

The iMarkup Annotation Tool: a Commer-

cial tool, viewed 13 September 2013,

http://www.bplogix.com/support/imarkup-client.aspx

PredictionofHumanPersonalityTraitsFromAnnotationActivities

269