Finding Insights between Active Aging Variables: Towards a

Data Mining Approach

María-Inés Acosta-Urigüen

, Priscila Cedillo

1,2 b

, Marcos Orellana

, Alexandra Bueno

Juan-Fernando Lima

and Daniela Prado

Laboratorio de Investigación y Desarrollo en Informática – LIDI, Universidad del Azuay, Cuenca, Ecuador

Universidad de Cuenca, Cuenca, Ecuador

Keywords: Active Aging, Data Mining, Cognitive Evaluation.

Abstract: Several proposals on active aging have been addressed within the psychological field, conceptualizing it

satisfactorily as a perspective of aging. Those proposals generate indicators that assess the level of physical

health, psychological wellbeing, adequate social adaptation. Physical, cognitive, and functional faculties,

interpersonal relationships, and productive activities have been evaluated. Although several technological

approaches have been proposed to promote active aging, they have not included a deep understanding of the

results obtained from solution implementations. Then, this paper presents the first step towards an approach

that uses variables proposed by active aging models (e.g., health, cognition, activity, affection, fitness aspects)

to generate knowledge through patterns. These patterns are identified using data obtained through several

instruments (i.e., psychological evaluations, health studies, and human experts' contributions). Thus, selecting

those variables and evaluating them as future models is necessary. Domain experts perform this evaluation.

The evaluation of this proposal has been completed with participants belonging to the health area through a

case study. This evaluation generates input data for engineers to apply data mining techniques to reveal

strategic knowledge. Finally, from the psychologist's point of view, the results showed that the contribution

results are appropriate for achieving healthy aging indicators.

1 INTRODUCTION

When the old age concept was analyzed, it was only

related to illness, memory problems, senility,

dementia, poverty, and depression (Lupien & Wan,

2004). However, the concept of old age can be

addressed, considering the quality of life as a relevant

factor contributing to it. The World Health

Organization (WHO), in 2015, presented a World

Report on Aging and Health (World Health

Organization, 2015). This report covers aging as the

sum of several changes. Biologically, aging is

associated with accumulating a wide variety of

molecular and cellular damages that gradually reduce

physiological reserves, increasing the risk of many

https://orcid.org/0000-0003-4865-2983

https://orcid.org/0000-0002-6787-0655

https://orcid.org/0000-0002-3671-9362

https://orcid.org/0000-0001-7188-1210

https://orcid.org/0000-0003-3500-3968

https://orcid.org/0000-0003-1241-1782

diseases and generally decreasing the individual's

capacity (World Health Organization, 2015).

The WHO considers three trajectories of healthy

aging: a) a period of relatively high and stable

capacity, b) a period of diminished capacity, and c) a

significant loss of abilities. It also identifies that there

are different ways to quantify active aging. Still, they

all keep the exact purpose of promoting and

maintaining intrinsic capacity. People with reduced

functional capacity can continue to carry out activities

that are important to them (World Health

Organization, 2015). The trajectory does not depend

on chronological age and is not uniform among

individuals.

268

Acosta-Urigüen, M., Cedillo, P., Orellana, M., Bueno, A., Lima, J. and Prado, D.

Finding Insights between Active Aging Variables: Towards a Data Mining Approach.

DOI: 10.5220/0011068100003188

In Proceedings of the 8th International Conference on Information and Communication Technologies for Ageing Well and e-Health (ICT4AWE 2022), pages 268-275

ISBN: 978-989-758-566-1; ISSN: 2184-4984

Worldwide, the concept of active aging has

gained interest. For example, the European Union has

focused on transitioning from the perception that

older adults are only recipients of retirement to a

practical orientation. They are active subjects at the

family, community, work, educational level, among

others (Walker & Maltby, 2012). Other clear

examples are observed in China, Japan, and South

Korea, where governments have proposed an

integrated, composite index for measuring the

contribution of older people to society, their

communities, and their families (Um et al., 2019).

Here, the generation of datasets and the

application of various techniques becomes an

attractive option to evaluate and interpret the concept

of active aging among a wide range of variables and

their combinations (Nayak, Buys, & Lovie-Kitchins,

2006). Also, the prediction, detection of outliers

(anomaly detection), clustering, and decision making

are some of the most used techniques belonging to the

data mining field (Gachet Páez et al., 2018; Moreira

& Namen, 2018).

Then, this paper proposes the first step towards an

approach that uses variables proposed by active aging

models (e.g., health, cognition, activity, affection,

fitness aspects) generates knowledge using patterns.

These patterns are identified using data obtained

through several instruments (i.e., psychological

evaluations, health studies, and human experts'

contributions). Thus, selecting those variables and

evaluating them as future models is necessary.

Domain experts perform this evaluation. The review

of this proposal has been completed with participants

belonging to the health area through a case study.

This evaluation generates input data for engineers to

apply data mining techniques to reveal strategic

knowledge. Finally, from the psychologist's point of

view, the results showed that the contribution results

are appropriate for achieving healthy aging

indicators.

Finally, this study is organized as follows: first,

the background and related work on applying data

mining in psychology are described; next, the

proposed approach which helps to characterize the

active aging; finally, conclusions and future work are

offered.

2 BACKGROUND

Even though few studies related to data mining

techniques applied in active aging, some have shown

promissory results. For example, Nayak, Buys, &

Lovie-Kitchins (2006) proposed the use of predictive

models to identify which variables need to be related,

from the groups labeled as work, learning, social,

spiritual, emotional, health, home, life events, and

demographics, contribute to achieving a positive

active aging score. Preprocessing techniques include

the value transformation removal of empty fields.

Then, the k-means clustering algorithm was used to

split the dataset into seven distinct, overlapped

clusters. They performed an association analysis and

rules to determine the meaning of the correlation of

variables.

The recognition of early variables of successful

aging can predict long-term survival. The study

presented by Swindell et al. (2010) was based on a

dataset taken from 4,097 women in the United States

of America, with several variables (i.e. demographic,

cognitive, familiar, medical variables). Their study

proposed a predictive model based on data mining

techniques to look for combinations of variables that

predict long-term survival. The result was a

composition of a 13-variable model.

A system that implements a big data approach

was presented to use bio-signal sensors and machine-

learning algorithms for recommendations (Gachet

Páez et al., 2018). It obtains data from wearable

sensors, and prediction, detecting outliers, clustering,

and decision making were applied for prediction.

Consequently, although there are several studies,

none considers applying data mining techniques to

propose models to improve that characterization.

Neither do they consider an expert validation

performed by psychologists? Therefore, this study

aims to provide a method that allows user data to

describe insights that help health personnel give the

best advice to their patients to reach active aging.

3 PROMOTING THE ACTIVE

AGING BASED ON DATA

SCIENCE

The Cross-Industry Standard for Data Mining

(CRISP-DM) proposes five high-level processes, that

contribute to carrying out a data science project. Each

step is described briefly as follows: 1) Business

understanding: it is focused on uncovering essential

factors, but in this case, it will be named background

understanding, 2) Data understanding, which

describes the acquisition of data listed in the project

resources, 3) Data preparation, which seeks to obtain

a specific dataset to be used as input during the entire

process, 4) Modeling, which is focused on selecting

the proper technique to be used, and finally, 5)

Finding Insights between Active Aging Variables: Towards a Data Mining Approach

269

Evaluation, which allows to accept or decline de the

generality of the model (Chapman et al., 2000).

Following these steps makes it feasible to reach good

results related to promoting active aging in patients.

All the conclusions and found knowledge are

reliable since it is based on data that have been

collected and can be equivalent to a domain expert

opinion; it is because data are a good source of

experience.

3.1 Background Understanding

Several theoretical perspectives have emerged that

differentiate between the "good" ways of aging in

recent years. At the end of the 60s of the 20th

centuries, various sociological perspectives appeared

that defined these ways of aging. Models of aging

based on exact psychological characteristics have

now been developed. Although these models have a

tremendous everyday basis, each of them represents a

different perspective or way of aging successfully.

Figure 1 shows how using models can identify the

variables useful for domain experts (i.e., health

personnel), which provide insights to improve

treatments and promote the well aging of their

patients. Those variables can be categorized to obtain

knowledge easier to manage depending on relevant

topics of preference for each health personnel.

Figure 1: Data comprehension pool.

3.1.1 Successful Aging

Successful aging refers to "a high level of physical

health, psychological wellbeing and adequate

adaptation" (Nadler et al., 1997). Successful aging

allows the person to reach an advanced age in full use

of their physical, cognitive, and functional faculties.

The Rowe and Kahn model was born and proposed to

include three main components for the definition of

successful aging: a) low probability of disease or

disability; b) high functional capacity at the cognitive

and physical level; and c) active commitment to life

(Stowe & Cooney, 2015).

3.1.2 Optimal Aging

One of the first proposals, presented by (Baltes &

Baltes 2010), indicates a development model

throughout the entire life cycle with the appropriate

interaction between three main elements: selection,

optimization, and compensation, adjusting their goals

and objectives to their vulnerability and the reduced

reserve capacity, typical of their age.

Brummel-Smith (2007b) points out that this form

of optimal aging should be considered a

biopsychosocial model. Happiness, an active social

life, or financial solvency contribute to healthy aging

in the psychosocial area. Finally, social support

facilitates coping and adaptation to changes and

protects individuals from stress-induced pathologies.

3.2 Data Comprehension

To accomplish this phase, the understanding of the

below phase is essential. The fields from the proposed

models are extracted and synthesized into one

collection of topics (see Figure 1).

To construct the topics and related variables, the

proposal of Lak et al. (2020) was considered. It

presents an iterative systematic review in the field of

active aging. The authors identified several studies

and summarized the most relevant variables into

different categories; these variables were included in

other models of active aging explained in section 2.

The considered topics are personal information,

personal behaviour, land use, social process,

policymaking, mental and physical health, and social

health.

Although there is no single model to measure

effective aging, the literature indicates that the tools

used by the authors are constituted in its measurement

based on the Likert scale (Echauri et al., 2013). The

Likert scale is an additive scale with an ordinal level

compound of a series of items to measure the subject's

reaction (Namakforoosh, 2002). Data mining can

play an important role by having quantitative data,

understanding the data, and finding clues by mining

information from surveys and tests (Koufakou et al.,

2016).

3.3 Data Preparation

The best way to avoid untrusted results data

considered in the study must be cleaned before using

data in the Modeling stage. The main activities of data

preparation are shown in Figure 2, and each activity

is part of the flow until to get a proper input dataset.

Identifying variables

Categorizing variables

Model 1..n

out

Active aging

variables

out

Topics

ICT4AWE 2022 - 8th International Conference on Information and Communication Technologies for Ageing Well and e-Health

270

Figure 2: Preprocessing the input dataset.

3.3.1 Likert Scale

This paper expects multiple fields based on the Likert

scale due to the input data. This point reflects the

available ways to deal with this kind of data. Likert-

Scale is commonly used as comfortable feedback, and

it is categorized as an ordinal data type. Also,

machine learning models evaluate and try to generate

the best goodness in their results (Kandasamy et al.,

2020); conversely, it is possible to deal with these

data only as ordinal data types and convert them to

their relative ordinal numbers.

3.3.2 Missing Values

No matter the type of data, it could be numerical or

not, discrete or continuous. If there are missing fields,

the Neuropsychologists experts prefer to backup up

and drop the records and all the linked records to

avoid the noise in the results (Nayak, Buys, & Lovie-

Kitchins. Literature such as Swindell et al. (2010)

shows that its variables are continuous and

categorical, proposing to fill in missing data with the

"Imputation approach" method. The missing data are

imputed based on the average value between the k =

20. In this stage, the choice about missing data

depends on the amount of data collected and the

opinion of experts in the domain (Enders, 2010).

3.3.3 Data Discretization

The variables are transformed into discrete data to

standardize continuous data inputs, establishing

ranges that cover data to be changed. As mentioned

in Nayak et al. (2006), one of the standard data to be

transformed is peoples' ages, who are part of the

group to be analyzed. According to the Research of

Adult Learning and Development, to manage these

age groups, a path is to transform ages to standardize

in groups as Childhood, Adolescence, Emerging

adulthood, Average adulthood, and Late adulthood,

according to the Research of Adult Learning and

Development (Smith & Nancy DeFrates-Densch,

2008). In the same way, Ethnicity can be categorized

according to Nerenz et al. (2009).

3.3.4 Data Cleaning

In data containing words or phrases, it is necessary to

identify words with values that will not contribute to

the data mining process. Instead, it will make the

process difficult. The typical stage includes

frequently repeated words in the writings, such as

articles or propositions (Moreira & Namen, 2018).

3.4 Modelling

Due to the diversity of data types, choosing the

appropriate technique to guarantee that data is

processed and showing the proper variables that

characterize active aging is the main interest of this

stage. Several approaches determine the influence of

variables as "Select by weights" that selects only

those whose weights satisfy a criterion concerning the

input weights (Malik & Mishra, 2014). However, we

seek to divide a group of interest into subgroups

(clusters). Therefore, an unsupervised method is a

proper approach.

The application of unsupervised methods has

been used in similar works as (Pal & Pal, 2013) where

evidence the predictive value of different measures of

cognition, based on clusters was found that girls with

the high socioeconomic status trend to higher

academic achievement in science stream, and boys

with low socioeconomic status had trends higher

academic achievement in general. Clustering on

social domains has been widely used as an

unsupervised technique for human activity

recognition (Ariza Colpas et al., 2020).

Clustering is essential for the Knowledge

Discovery tool because the goodness of data is

proportional to the fulfilment of its purpose. The

technique creates groups mutually exclusive based on

three possible conditions: a) Defining the maximum

cluster distance and minimizing it, b) Compute the

sum of averages of the distances and minimizing it,

and c) Compute the total cluster distance and

minimize it (Pandove et al., 2018).

Once the problem is apparent, and it is clear how

clustering can solve it, it is necessary to select the

proper technique among the numerous approaches

such as Hierarchical, K-means, Random sampling,

Randomize search, Condensation based, Density-

Based, Grid-based, Probabilistic Model-based, and

Clustering Graphs and Network Data (Pandove et al.,

2018). The requirement of this work defines two

segments: proper or no proper values for active aging

and based on this. K-means application is proposed to

include the number of groups before executing the

technique.

Finding Insights between Active Aging Variables: Towards a Data Mining Approach

271

Figure 3: Expected results after clustering.

According to (Kandasamy et al., 2020), the behavior

of the responses could be better understood since it

ensures that the most significant disadvantage of the

Likert scale is losing information and introducing

distortion to data. Therefore, applying clusters to

segment groups could be advantageous to find these

variables in the same group of people. The cluster

method that sticks to the data set of our characteristic

for effective aging is a proprietary method described

based on clustering to find the relationship of the

variables.

3.5 Evaluation

Due to the application of an unsupervised technique

done in the above section, the evaluation of the results

must be verified to verify what kind of variables

define active aging. Being clustering the method used

to analyze data, the proper way to show how grouped

data is related is through multidimensional graphs.

Generally, 2D or 3D dimensional (Kandogan, 2001)

display the groups. Therefore, an analysis of the

clusters is required to evaluate the goodness of the

clustering results.

Visual Analytics: The novel strategy of

incorporating human capabilities to describe the

behavior. The popular visualization methods used

are: Bar charts, Line charts, Pie charts, and Scatter

plots. Scatter plot is the standard method to describe

the clustering application, including 2D or 3D axes on

the cartesian, the human can evaluate according to the

dispersion of data. However, this is hard to assess and

validate if clusters are very closed (Chen et al., 2015).

4 CASE STUDY

This section presents the case study, with all the

activities proposed by Runeson et al. (2012). The

results of this evaluation represent an essential input

for data engineers; who analyze factors associated

with active aging in the proposed methodology.

Experts in active aging evaluate the inputs used in

the proposed methodology. It follows the method

proposed by Runeson et al. (2012). The activities to

be tracked are 1) design, 2) preparation for data

collection, 3) collecting evidence, 4) analysis of

collected data and reporting, and 5) threats of validity

analysis.

4.1 Design

A degree of agreement among the experts is obtained

from this evaluation regarding the variables

considered for healthy aging. Besides, this case study

aims to evaluate health personnel's perceptions

regarding the usefulness of the results derived from

artificial intelligence in the characterization of

healthy aging.

The evaluation's objectives and scope were

established with Goal Question Metric (GQM)

approach proposed by Basili et al. (Basili et al., 1994),

who offers a paradigm that defines the evaluations'

scope and objectives. The proposed GQM scheme

follows the scheme of a) The evaluation analyses the

inputs for the methodology proposed by the data

engineers; b) What is the purpose of the objective

measures the agreement among health experts on the

variables considered in the characterization of active

aging regarding the usefulness of the information

resulting from the methodology; c) From the point of

view of Clinical psychologists; and d) In the health

context where this study is carried out.

In this context, the research questions are: What

variables are considered important for characterizing

active aging for two health personnel, and what the

perception of health personnel on the usefulness of a

methodology that allows selecting active aging

variables collected utilizing artificial intelligence is.

According to Runeson et al. (Runeson et al.,

2012) recommendations, this case study method is

holistic-multiple, and the units of analysis are

presented in Figure 4.

Figure 4: Holistic-multiple method.

4.2 Preparation for Data Collection

Two surveys have been designed to achieve the

objectives of this case study. For Context 1, a form

(see https://n9.cl/iwle) has been made based on Lak,

Rashidghalam, Myint, and Bradaran (Lak et al.,

ICT4AWE 2022 - 8th International Conference on Information and Communication Technologies for Ageing Well and e-Health

272

2020), who propose a list of active aging

characteristics based on the study of related work.

The purpose of this first form is to reach a consensus

of experts in the gerontological health area of the

variables that should be considered for the

characterization of healthy aging through data

mining.

For Context 2, a form has been designed based on

the technology assessment model (TAM) proposed

by Davis (Davis, 1985). On this occasion, only the

constructs of the Perceived Utility (PU) and the Intent

to Use (UIT) in the future are analysed, specifically

of the final products of the methodology. Data mining

experts have designed the form with its respective

explanation. As shown in the following URL:

https://n9.cl/qcmb4, this questionnaire uses a 5-point

Likert scale.

4.3 Collecting Evidence

Both questionnaires were presented to two Clinical

psychologists with experience in the gerontological

area.

4.4 Data Analysis and Results

Reporting

By analyzing the results, it is found that they allow

answering the case study questions. In Case 1, Fleiss'

Kappa is used. It is a statistical measure for assessing

the reliability of agreement between a fixed number

of raters when assigning categorical ratings to several

items or classifying items; the action is scored

between 0 and 1 (0 means low agreement, and one

refers to a high deal). Fleiss' Kappa is used to validate

the process of inclusion/exclusion of variables

presented in Appendix 1 (see https://n9.cl/iwle).

Finally, the selection of each reviewer was checked,

and the discrepancies were resolved with consensus.

For the two raters, Fleiss's Kappa for agreement

on inclusion in the active aging characteristics was

0.82. Landis and Koch (Landis & Koch 1977) provide

a table to interpret the values, and values between

0.81 and 1.00 are considered almost perfect.

Furthermore, Fig. 5 presents the degree of agreement

by dimension. The open public space, housing, and

cultural environment reaches a fair deal, the social

climate qualifies a substantial agreement, and the rest

obtains an almost perfect agreement.

The average of the responses obtained for the two

TAM constructs analysed was calculated (see Fig. 6).

It is concluded that clinical psychologists mention

that this technological contribution can reduce the

time and effort of characterizing active aging in older

adults. Also, the participants recall that it is a valuable

input since it will allow an excellent characterization

of the study variable to develop future primary,

secondary, or tertiary intervention plans with the

elderly.

Figure 5: Fleiss Kappa measures per dimension.

Figure 6: Results of the case study - Clinical psychologist

perceptions.

4.5 Threats of Validity

Validity is measured analysing four categories:

construct, internal, and external validity, and

reliability.

Construct validity focuses on the relation between

the theory behind the study and the practical

experience. Ensure that the operational measures

studied represented what the researchers had planned

to investigate and what they were investigating. For

Context 1, a previous work by Sutano was used, who

proposed a list of variables derived from a rigorous

systematic review. Also, the Fleiss Kappa nominal

scale was used (Fleiss, 1971). It provides the degree

of experts agree to not fall into subjectivities.

Furthermore, in Context 2, a validated questionnaire

was used whose Cronbach's alpha is ideal. Thus,

constructs were interpreted in the same way by the

researcher and the interviewees.

Internal validity depends on how the participants

are selected. In this study, educational and

professional experience and the participant's

knowledge about the data mining field could

influence the responses and perceptions when using

the proposed solution. To mitigate this threat, the

selected participants have a similar professional

profile.

Due to the COVID-19 pandemic, the external

validity considered that the access to numerous

groups was restricted. Thus, selecting the sample of

individuals who participated was made at

Public open spa ce

Housing

Cultural environme nt

Social environment

0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

Economic environment

Good governance

Physical health

Mental health

Social health

Personal characteristics

Land use

Access

Physical form

City image

Finding Insights between Active Aging Variables: Towards a Data Mining Approach

273

convenience; for this reason, the results have to be

analysed carefully because they are not generalizable

to the population.

From the two forms design to the analysis of

results, the reliability considers how the evidence

chain was carried out to respect the data's literality.

Moreover, the qualitative responses were quantified

using a Likert scale to avoid introducing

interpretation bias or, failing that, the participants

mentioned textually.

5 CONCLUSIONS AND

FURTHER WORK

This paper allows the determination of variables'

values to determine active aging. Data mining allows

identifying among variables that are strongly

associated with the topic. Data recollected from

different sources in the psychological tests as mental,

physical, social, policy health, and personal behavior

(Fernandez-Ballesteros, 2011), these variables are

matched with variables of models proposed by the

WHO and Neuropsychologists (Nayak, Buys, &

Lovie-Kitchin, 2006).

The multiple models for measuring and

evaluating individuals' active aging have allowed

creating this framework to identify the appropriate

methods of active aging. Moreover, the proper

techniques to analyze them into each data mining

process: LikertSvm for Likert scale values, listwise

deletion for missing values, standardized data

discretization for sociodemographic variables

according to their categorization in health care.

Then, according to the input data and literature, a

clustering technique is proper to evaluate the groups

of active aging in the next stage of modeling.

However, the data scientists have to perform a

performance evaluation of clusters using metrics or

visual analytic analysis to get the best precision in

splitting groups.

Due to it not being a standardized technique o

evaluation to identify active aging, it is impossible to

classify the people in the two groups who have

successful aging among those who do not. Thus, the

number of variables to consider is significant, so

determining which are related is essential to open a

path to active aging.

After implementing the proposed framework and

getting results, we seek to report results after applying

different data science techniques.

ACKNOWLEDGMENTS

The authors wish to thank the Vice-Rector for

Research of the University of Azuay for the financial

and academic support and all the staff of the

Laboratory for Research and Development in

Informatics (LIDI), and the Department of Computer

Science of Universidad de Cuenca.

REFERENCES

Ariza Colpas, P., Vicario, E., De-La-Hoz-Franco, E.,

Pineres-Melo, M., Oviedo-Carrascal, A., & Patara, F.

(2020). Unsupervised Human Activity Recognition

Using the Clustering Approach: A Review. Sensors,

20(9), 2702. https://doi.org/10.3390/s20092702

Baltes, P. B., & Baltes, M. M. (2010). 成功衰老的心理学

观点：有偿的选择性优化模型. In Successful aging.

Basili, V. R., Caldiera, G., & Rombach, H. D. (1994). The

goal question metric approach. Encyclopedia of

Software Engineering, 2, 528–532.

https://doi.org/10.1.1.104.8626

Brummel-Smith, K. (2007). Optimal aging, part II:

Evidence-based practical steps to achieve it.

Chapman, P., Clinton, J., Kerber, R., Khabaza, T., Reinartz,

T., Shearer, C., Wirth, R., & others. (2000). CRISP-DM

1.0: Step-by-step data mining guide. SPSS Inc, 9, 13.

Chen, W., Guo, F., & Wang, F. (2015). A Survey of Traffic

Data Visualization. IEEE Transactions on Intelligent

Transportation Systems, 16(6), 2970–2984.

https://doi.org/10.1109/TITS.2015.2436897

Davis, F. (1985). A technology acceptance model for

empirically testing new end-user information systems:

Theory and results [Massachusetts Institute of

Technology]. In Massachusetts Institute of Technology.

https://doi.org/oclc/56932490

Enders, C. K. (2010). Applied missing data analysis.

Guilford press.

Fabila Echauri, A. M., Minami, H., & Izquierdo Sandoval,

M. J. (2013). La Escala de Likert en la evaluación

docente : acercamiento a sus características y principios

metodológicos. Textos y Contextos, 31–40.

Fernandez-Ballesteros, R. (2011). Positive ageing.

Objective, subjective, and combined outcomes. EJAP

(Test), 7(1), 22–30.

Fleiss, J. L. (1971). Measuring nominal scale agreement

among many raters. Psychological Bulletin, 76(5), 378–

382. https://doi.org/10.1037/h0031619

Gachet Páez, D., de Buenaga Rodríguez, M., Puertas Sánz,

E., Villalba, M. T., & Muñoz Gil, R. (2018). Healthy

and wellbeing activities’ promotion using a Big Data

approach. Health Informatics Journal, 24(2), 125–135.

https://doi.org/10.1177/1460458216660754

Kandasamy, I., Kandasamy, W. B. V., Obbineni, J. M., &

Smarandache, F. (2020). Indeterminate Likert scale:

feedback based on neutrosophy, its distance measures

ICT4AWE 2022 - 8th International Conference on Information and Communication Technologies for Ageing Well and e-Health

274

and clustering algorithm. Soft Computing, 24(10),

7459–7468.

Kandogan, E. (2001). Visualizing Multi-Dimensional

Clusters, Trends, and Outliers Using Star Coordinates.

Proceedings of the Seventh ACM SIGKDD

International Conference on Knowledge Discovery and

Data Mining, 107–116. https://doi.org/10.1145/

502512.502530

Koufakou, A., Gosselin, J., & Guo, D. (2016). Using data

mining to extract knowledge from student evaluation

comments in undergraduate courses. 2016

International Joint Conference on Neural Networks

(IJCNN), 3138–3142.

Lak, A., Rashidghalam, P., Myint, P. K., & Bradaran, H. R.

(2020). Comprehensive 5P framework for active aging

using the ecological approach: An iterative systematic

review. BMC Public Health, 20(1), 1–22.

https://doi.org/10.1186/s12889-019-8136-8

Landis, J. R., & Koch, G. G. (1977). The Measurement of

Observer Agreement for Categorical Data. Biometrics.

https://doi.org/10.2307/2529310

Lupien, S. J., & Wan, N. (2004). Successful ageing: From

cell to self. Philosophical Transactions of the Royal

Society B: Biological Sciences, 359(1449), 1413–1426.

https://doi.org/10.1098/rstb.2004.1516

Malik, H., & Mishra, S. (2014). Feature selection using

RapidMiner and classification through probabilistic

neural network for fault diagnostics of power

transformer. 2014 Annual IEEE India Conference

(INDICON), 1–6.

Moreira, L. B., & Namen, A. A. (2018). A hybrid data

mining model for diagnosis of patients with clinical

suspicion of dementia. Computer Methods and

Programs in Biomedicine, 165, 139–149.

https://doi.org/10.1016/j.cmpb.2018.08.016

Nadler, J. D., Damis, L. F., & Richardson, E. D. (1997).

Psychosocial Aspects of Aging.

Namakforoosh, M. N. (2002). Metodología de la

Investigación.

Nayak, R., Buys, L., & Lovie-Kitchin, J. (2006). Data

Mining in Conceptualising Active Ageing. In J. Li, S.

J. Simoff, P. J. Kennedy, P. Christen, & G. J. Williams

(Eds.), Data Mining and Analytics 2006: Proceedings

of the Fifth Australasian Data Mining Conference (pp.

39–46). Australian Computer Society Inc.

Nayak, R., Buys, L., & Lovie-Kitchins, J. (2006). Data

Mining in Conceptualising Active Ageing BT - Fifth

Australasian Data Mining Conference (AusDM2006).

61, 39–45. http://crpit.com/confpapers/CRPITV61

Nayak.pdf

Nerenz, D. R., McFadden, B., & Ulmer, C. (2009). Race,

Ethnicity, and Language Data: Standardization for

Health Care Quality Improvement. National

Academies Press.

Pal, A. K., & Pal, S. (2013). Evaluation of teacher’s

performance: a data mining approach. International

Journal of Computer Science and Mobile Computing,

2(12), 359–369.

Pandove, D., Goel, S., & Rani, R. (2018). Systematic

Review of Clustering High-Dimensional and Large

Datasets. ACM Trans. Knowl. Discov. Data, 12(2).

https://doi.org/10.1145/3132088

Runeson, P., Höst, M., Rainer, A., & Regnell, B. (2012).

Case Study Research in Software Engineering:

Guidelines and Examples. In Case Study Research in

Software Engineering: Guidelines and Examples.

https://doi.org/10.1002/9781118181034

Smith, M. C., & Nancy DeFrates-Densch, A. E. (2008).

Handbook of Research on Adult Learning and

Development. Taylor & Francis.

Stowe, J. D., & Cooney, T. M. (2015). Examining rowe and

kahn’s concept of successful aging: Importance of

taking a life course perspective. Gerontologist, 55(1),

43–50. https://doi.org/10.1093/geront/gnu055

Swindell, W. R., Ensrud, K. E., Cawthon, P. M., Cauley, J.

A., Cummings, S. R., & Miller, R. A. (2010). Indicators

of “healthy aging” in older women (65-69 years of age).

A data-mining approach based on prediction of long-

term survival. BMC Geriatrics, 10, 1–24.

https://doi.org/10.1186/1471-2318-10-55

Um, J., Zaidi, A., & Choi, S. J. (2019). Active Ageing Index

in Korea – Comparison with China and EU countries.

Asian Social Work and Policy Review, 13(1), 87–99.

https://doi.org/10.1111/aswp.12159

Walker, A., & Maltby, T. (2012). Active ageing: A strategic

policy solution to demographic ageing in the European

Union. International Journal of Social Welfare,

21(SUPPL.1), 1–10. https://doi.org/10.1111/j.1468-

2397.2012.00871.x

World Health Organization. (2015). Informe mundial sobre

el envejecimienot y la salud.

Finding Insights between Active Aging Variables: Towards a Data Mining Approach

275