Business Intelligence Reporting by Linguistic Summaries for Smart

Cities: A Case on Explaining Bicycle Sharing Patterns

Erika Min

arikov

1 a

, Galena Pisoni

2 b

, B

alint Moln

3 c

and Hanna Krist

ın Skaftadottir

4 d

Faculty of Economic Informatics, University of Economics in Bratislava, Bratislava, Slovakia

York Business School, York St. John University, Lord Mayor’s Walk, YO31 7EX, York, U.K.

otv

os Lor

and University, ELTE, IK P

azm

any P

eter 1/C, 1117, Budapest, Hungary

Department of Business, Bifrost University, Iceland

Keywords:

Linguistic Summaries, Smart Cities, Enterprise Architectures, Business Intelligence Reporting, Drill-Down

and Roll-up Summaries.

Abstract:

An increasing number of intelligent urban services rely on the use of Information and Communication Tech-

nologies (ICT). Data-driven approach is often considered for supporting sustainable cities, provided the per-

vasive nature of the Internet of Things (IoT) like sensors, and their capabilities to collect data for elaborating

to the cities. This paper focuses on an intelligent business reporting approach explaining the bicycle sharing

patterns by linguistic summaries in order to provide relevant insights for decision makers and citizens. We ex-

plored the developments in bicycle sharing stations in different periods of the day for months and seasons. The

business intelligence query operations of drill-down and roll-up are often used in data reporting and analysis.

In this work, these operations are realized by linguistic summaries. The main aim is to propose an approach for

analysis and visualization in an understandable and interpretable way for diverse user categories. Experiments

were conducted on the Dublin bicycle sharing data set. Finally, a way how cities can set in place the collection

of data coming from different sources, as well as relevant enterprise infrastructures and data analytic pipelines

for such service are discussed.

1 INTRODUCTION

The fast and undeniable growth of urban mobility ser-

vices put us in the front of new challenges. New sus-

tainable and on-demand mobility solutions have un-

preceded growth, producing therefore also lots of data

to work within this ﬁeld (Kong et al., 2020; Mayer-

Sch

onberger and Cukier, 2013; Provost and Fawcett,

2013). These developments require: on one hand

(i) adequate enterprise architectures and solutions to

properly transform and processes data coming from

multiple sources, and therefore provide the basis for

data analysis, and on the other hand (ii) adequate so-

lutions for data analysis and interpretations of data to

diverse respondents groups.

Smart cities should not only offer sustainable mo-

bility services, but also work towards their success-

https://orcid.org/0000-0002-4230-2109

https://orcid.org/0000-0002-3266-1773

https://orcid.org/0000-0001-5015-8883

https://orcid.org/0000-0001-5228-8294

ful implementation and optimization of resources al-

located to them (Lim et al., 2018; Torre-Bastida et al.,

2018). In this work, we tackle the problem of bicycle

sharing, and the explaining use of bicycle stations that

recently appear in different cities as a possible solu-

tion for the mobility problems (Midgley, 2009; Midg-

ley, 2011). Various contemporary questions / prob-

lems that cities should tackle for the mobility services

exists. For instance, which stations are usually very

busy? In which time of the day considered bicycle sta-

tions are full? Does the number of tourists inﬂuence

the availability of bicycles? Are bicycle stations less

busy during the summer? The answers to such ques-

tions support the local authority in improving services

by, for instance, adjusting capacities of the bicycle

stations. For this task, the relevant data should be col-

lected, analyzed and presented in an understandable

way for local authorities (to manage improvements -

adjust the capacity of the stations accordingly), cit-

izens (to understand changes), journalists (to report

developments), etc.

Digitization and data processing developments of

762

Mináriková, E., Pisoni, G., Molnár, B. and Skaftadottir, H.

Business Intelligence Reporting by Linguistic Summaries for Smart Cities: A Case on Explaining Bicycle Sharing Patterns.

DOI: 10.5220/0012748200003690

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 26th International Conference on Enterprise Information Systems (ICEIS 2024) - Volume 2, pages 762-768

ISBN: 978-989-758-692-7; ISSN: 2184-4992

the past decades have transformed the ﬁeld of smart

cities and urban mobility profoundly. Diverse data

sources are available and therefore can be handled for

improving informativeness and support decision mak-

ing. Internet of Things (IoT)/sensors, and information

systems of the transport enterprises are examples of

the mobility data.

One of the aspects of human reasoning and deci-

sion is searching for information that is not immedi-

ately seen in the collected data (Trillas, 2015), prefer-

ably in an understandable way. When we focus on

the interpretation of information from data, we should

bear in mind that diverse urban stakeholder categories

differ in the levels of the statistical and IT literacy

(Hudec et al., 2018). Thus, the information should

be digested and interpreted in the most suitable way,

e.g., by linguistic summaries.

In this contribution, we explore the concept known

as linguistic summaries (see, e.g., (Boran et al.,

2016)) to reveal, whether these summaries are beneﬁ-

cial and how we can improve business intelligence re-

porting with the linguistic summaries supporting the

usual business intelligence operations of drill-down

and roll-up.

From what discussed above, we set the following

research questions

• RQ1. How can relevant enterprise architectures

be set to collect and transform urban mobility data

coming from different sources?

• RQ2. Could short quantiﬁed sentences improve

explainability of data?

The article is organized as follows. Section 2

brieﬂy explains preliminaries of linguistic summaries

and data set used for experiments. Section 3 is de-

voted to the procedure, experiments and mining sum-

marized sentences from the data set, whereas Sec-

tion 4 is devoted to the enterprise architecture for sup-

porting summaries. Section 5 discusses obtained re-

sults, perspectives and future tasks. Finally, Section 6

concludes the article.

2 METHODS, METHODOLOGY

AND DATA SET

This section introduces linguistic summaries and data

set.

2.1 Preliminaries of Linguistics

Summaries

Linguistic summarization of data is a topic which oc-

cupies scientists and practitioners since Yager’s sem-

inal work (Yager, 1982). Linguistic summaries have

been improved and applied in diverse ﬁelds, e.g., (Bo-

ran et al., 2016; Smits et al., 2018; van der Heide

and Trivino, 2009; Wilbik et al., 2020). A summary

like: in the morning the bike station is very busy, or

the most of young citizens commute large distances to

ofﬁces is understood at ﬁrst glance. Linguistic sum-

maries have not been applied only to interpret data,

but also for revealing dependencies between data and

satellite images in smart cities images (Hudec et al.,

2020) among others.

Two main structures of the classic prototype forms

are so-called basic structure of linguistic summaries

(LS) and structure with restriction (Lesot et al., 2016).

The basic structure is Q records have S. Quantiﬁer

Q and summarizer S are usually formalized by fuzzy

sets.

The proportion (relative cardinality) of entities in

a data set X that fully and partially satisﬁes the sum-

marizer (predicate) S is (Yager, 1982)

(X) =

∑

i=1

) (1)

where n is the number of entities and the member-

ship function µ formalizes summarizer S. The validity

(truth value) of the summary is calculated as

(X) = µ

(X)) (2)

where the function µ

formalizes fuzzy relative quan-

tiﬁer Q for the summary.

The structure with restriction is Q R records have

S. The proportion of entities in data set that meet the

summarizer S and restriction R is (Yager, 1982)

(X) =

∑

i=1

t(µ

), µ

))

)

(3)

where the membership functions µ formalize summa-

rizer S and restriction R. The validity (truth value) of

the summary is calculated as (2).

2.2 Datsets

For our experiments we used the open

data set of the bicycle sharing in Dublin,

Ireland. This data set is accessible at

https://data.smartdublin.ie/dataset/dublinbikes-api.

The considered data set contains collected data re-

lated to the accessibility of bicycles on the bicycles

station points within the city of Dublin on ﬁve minute

interval for the year 2021. The following attributes

are available:

station id, date time, last updated, name, bike stands,

available bike stands, available bikes, status, address,

latitude, longitude.

Business Intelligence Reporting by Linguistic Summaries for Smart Cities: A Case on Explaining Bicycle Sharing Patterns

763

In the pre-processing steps the time was extracted

from the date time attribute and adjusted for the lin-

guistic terms explained later on. The availability of

bicycles was calculated as a ratio between the bicy-

cle stands capacity and and the available bicycles for

every collected record. The Python libraries numpy

pandas and datetime are used for prepossessing. All

calculations of summaries are also realized in Python.

3 SUMMARISING BICYCLE

SHARING DATA

This section focuses on summarizing data and report-

ing revealed patterns compatible to the usual business

intelligence queries.

3.1 Procedure

Generally, LSs provide validity of any summary

posed on data. However, the usual business intel-

ligence queries supports operations of roll up and

drill down (Kimball, 2011). The former gives a

global overview, e.g., for top managers or sharehold-

ers, whereas the latter provides details on lower hier-

archical level, e.g., for a region with a poor behavior

to see, which districts are the most problematic ones,

or whether poor behavior is in all districts.

In our procedure for summarising bicycle shar-

ing data, we focuses on summarizing developments

in stations by seasons. The roll up operation gives a

global overview of a station for citizens or journalists

for the entire year. The drill down operation is applied

for seasons without a recognized pattern. The aim is

to reveal, whether we can ﬁnd patterns on the months

or days levels.

3.2 Experiments

In order to proceed with revealing all relevant linguis-

tic summaries from the afore explained data set (Sec-

tion 2.2), we deﬁned fuzzy sets for linguistic terms

morning, around lunch, evening and night on the time

attribute as is shown in Figure 1. We emphasize, that

these fuzzy sets cover an usual vague separation of

these parts of the day. When in a particular city or re-

gion is a different meaning of these terms, fuzzy sets

can be adjusted accordingly.

The next linguistic variable is availability of bicy-

cle in stations. The number of bicycle stands in sta-

tions differ. Thus, instead of of number of bicycles we

adopted the proportion of available bicycles, which is

the ratio of the number of stands and available bicy-

cles. This ratio is fuzziﬁed into three fuzzy sets few,

Figure 1: Part of days fuzziﬁed into four fuzzy sets.

about half and most of bicycles are available. The

next required concept is an elastic quantiﬁer most of

to express sentence like the most of mornings (in the

considered time period) few bikes is available.

In the next step, we executed summaries for each

combination of time and bicycle availability and

select the most relevant quantiﬁed sentences. We

calculated summaries for two stations. For the other

stations, the calculations are straightforward. In order

to shorten sentences, we excluded quantiﬁer most

of, i.e., instead of the most of mornings few bikes is

available we write mornings is available few bikes.

The results (the sentences with the highest validity)

for stations ID(117) and ID(38) are as follows.

Season:01-04 (Winter)

Station ID(117)

morning is available few bikes t= 0.9748

around lunch is available few bikes t= 0.9875

evening is available few bikes t= 0.9852

night is available few bikes t= 0.9820

Season:01-04 (Winter)

Station ID(38)

morning is available around

half bikes t= 0.5023

around lunch is available few bikes t= 0.5277

evening is available around half bikes t= 0.5635

night is available around half bikes t= 0.5372

Season:07-10 (Summer)

Station ID(117)

morning is available few bikes t= 0.9592

around lunch is available few bikes t= 0.9740

evening is available few bikes t= 0.9666

night is available few bikes t= 0.9628

Season:07-10 (Summer)

Station ID(38)

morning is available many bikes t= 0.3975

around lunch is available few bikes t= 0.4227

evening is available around half bikes t= 0.4764

night is available many bikes t= 0.4494

Summaries revealed that the station ID(117) is

very busy during both seasons, whereas for station

ID(38) the dominant bicycles availability pattern does

SEC-SCIS 2024 - Special Session on Soft Computing in Ethicity and Smart Cities Services

764

not exist. Thus, we try by the drill down operation to

the months level to ﬁnd, whether we can recognize

patterns in respective months.

Drill down to months (Station ID 38 summer):

July

morning is available many bikes t= 0.4246

around lunch is available around half bikes t= 0.451

evening is available many bikes t= 0.5129

night is available many bikes t= 0.5189

No recognised dominant pattern, we continue

with drill down on days:

01.07.2021

morning is available around half bikes t= 0.0

morning is available few bikes t= 1.0

02.07.2021

morning is available around half bikes t= 0.8537

morning is available few bikes t= 0.0269

03.07.2021

morning is available many bikes t= 0.9944

morning is available around half bikes t= 0.0083

morning is available few bikes t= 0.0009

etc.

This drill down operation is realizable due to the

data availability for each minute.

For each day summaries recognised different be-

haviour. This is the reason, why with the aggregated

data no dominant pattern is recognised.

August

morning is available many bikes t= 0.4351

around lunch is available few bikes t= 0.4834

night is available many bikes t= 0.4671

September

morning is available few bikes t= 0.3718

around lunch is available few bikes t= 0.4597

evening is available around half bikes t= 0.5172

No signiﬁcant pattern for both months - another

drill down is required.

October

morning is available many bikes t= 0.5701

around lunch is available around half bikes t=

0.5968

evening is available many bikes t= 0.9583

night is available many bikes t= 1.0

In the last month, we recognized that many bicy-

cles are available in evening and night.

The opposite operation in roll up to the year 2021

for this station (ID38):

morning is available around half bikes t= 0.382

around lunch is available around half bikes t=

0.4699

evening is available around half bikes t= 0.4722

night is available around half bikes t= 0.4175

Again, not a dominating pattern is recognized. In

our experiments, we considered pattern to be domi-

nant when its validity is greater than or equal to 0.75.

For station ID117 there is not a signiﬁcant differ-

ence between the considered sessions. The revealed

patterns indicate that constantly only few bicycles is

available. This station is very busy during the entire

year. Thus, the increasing in capacity should be con-

sidered, or at least the focus of business intelligence

dashboard should be on this station.

Contrary, station ID38 is rarely busy. Thus, stands

can be reduced and moved to another station. The

most relevant sentence has validity slightly above 0.5.

4 ENTERPRISE ARCHITECTURE

FOR DATA COLLECTION AND

MINING SUMMARIES

Recently, the ﬁelds of smart mobility became one of

the primary sources of data through the applications

of various sensors and IoT-s. In the ecosystem of

smart cities and its smart mobility, the efﬁcient and

effective utilization of data become an essential is-

sue. Transformation of data collection from the sim-

ple substantiated IoT data to data that originate from

large-scale monitoring led to the requirement of dis-

ciplined data analytic. Data can be listed as a tradi-

tional data coming from the city operation, and op-

erational data of transport originated from different

devices. The latter data are usually unstructured and

heterogeneous; either we consider their structure or

their content. Data should be accompanied by meta-

data that describe the essential information about the

content and can be utilized to categorize and organize

data to exploit them for the advanced data mining.

When focusing on the bicycle sharing, it depends

on various aspects like weather, restrictions (e.g., pan-

demic or events). Such data are usually not available

in the data collected from sensors on the stations.

Previous research has shown that analysis of data

that originated from structured databases and exter-

nal sources of a clear relation to structured data is

Business Intelligence Reporting by Linguistic Summaries for Smart Cities: A Case on Explaining Bicycle Sharing Patterns

765

relatively straightforward with the use of data ware-

houses (Golfarelli and Rizzi, 2009; Kimball, 2011;

Kimball and Ross, 2013). In the bicycle sharing, such

data can be from the meteorological data sets explain-

ing weather conditions near each station, or whether

events appear near the stations. The essential function

of the data warehouse is to integrate data and trans-

form them into the conﬁrmed structure, relevant for

the domain of interest. In this case, the domain of the

urban sustainable mobility to support reporting and

interpreting for the diverse urban stakeholders groups.

In Section 3.2, linguistic summaries are applied

on the bicycle sharing data to reveal patterns. Bene-

ﬁcially, the Key Performance Indicator (KPI) for the

stakeholders should be expressed linguistically (Vais-

man and Zim

anyi, 2022). The reporting with linguis-

tic summaries gives a reliable, trustable, and easy-to-

understand overview of the actual state regarding the

stations (see Section 3.2). This interpretations provide

the basis for supporting effective decision-making for

the city (see Figure 2). In our case, KPI is expressed

by linguistic terms: few, about half and most of.

In order to generate the linguistic insights as afore

computed, the relevant ”data ingestion” framework,

or the way how the data are fed to the data ware-

house, should be set in place. Relevant levels of secu-

rity should be treated in line with existing regulation

(foremost GDPR) for data management.

The beneﬁt of implementing an approach as pro-

posed in this work, is that a local authority can, if of

interest, extend the reporting of the collected data by

the other internal and external data and adjust report-

ing to diverse user categories. For instance, citizens

and users with disabilities beneﬁt of LS, which pro-

vide an easy way to communicate mined patterns. For

advanced users, like trafﬁc experts, external data cov-

ering, e.g., trafﬁc densities of cars, weather conditions

and organized events can reveal how these parame-

ters inﬂuence bicycle sharing in the affected stations.

With an approach as is under the development in this

work, local authorities would have a better overview

of cycling in the city. Next, suited LSs can be adopted

for informing cyclists about their riding behaviour in

comparison to the average values, for instance. How-

ever, a summary of structure your length of ride is

around average is not informative enough. This sum-

mary is the same for person who use bicycle every

working day (but not during weekends) and for some-

one who rarely cycle on working days, but is a heavy

bicycle user during weekends. The better option are

quantiﬁed linguistic summaries like few days your

length of rides is slightly under average, and about

half of days your length of rides is about average.

5 DISCUSSION

The proposed approach has a signiﬁcant applicability

potential. This holds especially when informing di-

verse urban stakeholders groups (including disabled

citizens) should be realized by a robust and compact

approach.

In our data intense society, we face the problems

of explaining mined patterns (Smits et al., 2018). A

promising way is by linguistic summaries. In order to

contribute, we raised two research questions. The ﬁrst

research question is How can relevant enterprise ar-

chitectures be set to collect and transform urban data

coming from different sources? The answer is that the

processing data by the linguistics summaries requires

a clear data warehouse model for storing data coming

from diverse sources. Such data are relevant for ad-

vanced explaining summaries and revealed patterns.

The second research question is Could short quan-

tiﬁed sentences improve explainability of data? The

answer depends on the structure of the sentences. If

it is a basic structure of summary (1), then histograms

and the other charts are suitable when visual attention

is not disturbed (i.e., it should not be focused else-

where) (Arguelles and Trivi

no., 2013), or summary

is not for visually impaired citizens. Regarding, the

summary with restriction, the situation is different. It

is more convenient to express it by sentences like in

Section 3.2 than on series of graphs covering more at-

tributes. Anyway, linguistically summarized sentence

is convenient for all urban stakeholders groups.

The next perspective is merging summaries with

maps. Either by using map features in summa-

rized sentences (Hudec et al., 2020), like whether a

particular proportion of objects inﬂuences the other

attributes, or interpreting summarized sentences on

maps. The positions of two stations used in exper-

iments (Section 3.2) are shown on map in Figure 3.

The very busy station is located near the river and

canal.

The main problem of mining all relevant summa-

rized sentences is the computational cost, when we

consider all stations and all possible summaries in a

large city. In the future work, we focus our work on

the optimisation in this direction.

6 CONCLUSIONS

A smart and sustainable mobility should rely on a

heavier use of bicycles. One possibility is by the bi-

cycle sharing concept. This work has shed light on

summarizing the bicycle sharing patterns of the bicy-

cle stations.

SEC-SCIS 2024 - Special Session on Soft Computing in Ethicity and Smart Cities Services

766

Figure 2: Sources of data and usage, ﬁgure composed by the authors, the creation of the ﬁgure was inspired by (Kimball,

2011; Mayer-Sch

onberger and Cukier, 2013; Provost and Fawcett, 2013; Kong et al., 2020).

Figure 3: Location of the two stands used in experiments.

In business intelligence reporting a well-known

way is by drill down and roll up query operations

(e.g., in MDX query language). In this work, we cre-

ated these operations by linguistic summaries. Next,

we demonstrated summaries (including drill down

and roll up) on the real world open dataset coming

from the bicycle sharing service in Dublin, Ireland.

In this paper, we also proposed an architecture

design required for collecting and transforming data

coming from different sources as a support for reveal-

ing and interpreting patterns by linguistic summaries.

It is an explainable way how to provide insights into

the bicycle sharing data, a relevant support for deci-

sion makers, local authorities and citizens.

This position work requires further investigation

in the ﬁeld of effective data collection and integration

Business Intelligence Reporting by Linguistic Summaries for Smart Cities: A Case on Explaining Bicycle Sharing Patterns

767

due to diversity of required data, as well as in the op-

timization of mining relevant patterns of bicycle shar-

ing by linguistic summaries. Linguistic summariza-

tion copes with the high computational demand due

to a larger number of data expressing bicycle sharing

stations and variety in summarized sentences.

ACKNOWLEDGEMENTS

This research was supported the Thematic Excellence

Programme TKP2021-NVA-29 (National Challenges

Subprogramme) funding scheme, the COST Action

CA19130 - ”Fintech and Artiﬁcial Intelligence in Fi-

nance Towards a transparent ﬁnancial industry” (Fi-

nAI), and the FBR-PDI-021 - The cooperation in the

ﬁelds of business intelligence and artiﬁcial intelli-

gence for science and education - BICISEDU.

REFERENCES

Arguelles, L. and Trivi

no., G. (2013). I-struve: Auto-

matic linguistic descriptions of visual double stars.

Engineering Applications of Artiﬁcial Intelligence,

26:2083–2092.

Boran, F., Akay, D., and Yager, R. (2016). An overview of

methods for linguistic summarization with fuzzy sets.

Expert Systems with Applications, 61:356–377.

Golfarelli, M. and Rizzi, S. (2009). A survey on tempo-

ral data warehousing. International Journal of Data

Warehousing and Mining (IJDWM), 5(1):1–17.

Hudec, M., Bedn

arov

a, E., and Holzinger, A. (2018). Aug-

menting statistical data dissemination by short quanti-

ﬁed sentences of natural language. Journal of Ofﬁcial

Statistics, 34:981–1010.

Hudec, M., Vu

ceti

c, M., and

Cerm

akov

a, I. (2020). The

synergy of linguistic summaries, fuzzy functional de-

pendencies and land coverings for augmenting infor-

mativeness in smart cities. In 28th Telecommunica-

tions forum TELFOR 2020. IEEE.

Kimball, R. (2011). The evolving role of the enterprise data

warehouse in the era of big data analytics. Whitepaper,

Kimball Group, April.

Kimball, R. and Ross, M. (2013). The data warehouse

toolkit: The deﬁnitive guide to dimensional modeling

(3rd ed.). Wiley, New York.

Kong, L., Liu, Z., and Wu, J. (2020). A systematic review of

big data-based urban sustainability research: State-of-

the-science and future directions. Journal of Cleaner

Production, 273:123142.

Lesot, M.-J., Moyse, G., and Bouchon-Meunier, B. (2016).

Interpretability of fuzzy linguistic summaries. Fuzzy

Sets and Systems, 292:307–317.

Lim, C., Kim, K.-J., and Maglio, P. P. (2018). Smart cities

with big data: Reference models, challenges, and con-

siderations. Cities, 82:86–99.

Mayer-Sch

onberger, V. and Cukier, K. (2013). Big data: A

revolution that will transform how we live, work, and

think. Houghton Mifﬂin Harcourt.

Midgley, P. (2009). The role of smart bike-sharing systems

in urban mobility. Journeys, 2(1):23–31.

Midgley, P. (2011). Bicycle-sharing schemes: enhancing

sustainable mobility in urban areas. United Nations,

Department of Economic and Social Affairs, 8:1–12.

Provost, F. and Fawcett, T. (2013). Data science and its rela-

tionship to big data and data-driven decision making.

Big data, 1(1):51–59.

Smits, G., Nerzic, P., Pivert, O., and Lesot, M. (2018). Ef-

ﬁcient generation of reliable estimated linguistic sum-

maries. In 2018 IEEE International Conference on

Fuzzy Systems (FUZZ–IEEE). IEEE.

Torre-Bastida, A. I., Del Ser, J., La

na, I., Ilardia, M., Bil-

bao, M. N., and Campos-Cordob

es, S. (2018). Big

data for transportation and mobility: recent advances,

trends and challenges. IET Intelligent Transport Sys-

tems, 12(8):742–755.

Trillas, E. (2015). An algebraic model of reasoning to

support zadeh’s cww. In Kacprzyk, J. and Pedrycz,

W., editors, Handbook of Computational Intelligence,

pages 249–267. Springer, Berlin Heidelberg.

Vaisman, A. and Zim

anyi, E. (2022). Data Warehouse Sys-

tems – Design and Implementation. Springer-Verlag,

Berlin Heidelberg.

van der Heide, A. and Trivino, G. (2009). Automatically

generated linguistic summaries of energy consump-

tion data. In Ninth International Conference on In-

telligent Systems Design and Applications (ISDA ’09).

IEEE.

Wilbik, A., Barreto, D., and Backus, G. (2020). On rele-

vance of linguistic summaries – a case study from the

agro–food domain. In 8th International Conference

on Information Processing and Management of Un-

certainty in Knowledge–Based Systems (IPMU 2020).

Springer.

Yager, R. (1982). A new approach to the summarization of

data. Information Sciences, 28(1):69–86.

SEC-SCIS 2024 - Special Session on Soft Computing in Ethicity and Smart Cities Services

768