Improving Czech Digital Government Based on Quantified Maturity

Model of Enterprise Architecture

Martin Rod

1,2 a

and Jiri Vomlel

Faculty of Management, Prague University of Economics and Business, Jarosovksa 1117/II,

Jindrichuv Hradec, 377 01, Czechia

Institute of Information Theory and Automation, Czech Academy of Sciences, Pod Vodarenskou vezi 4,

Prague, 182 00, Czechia

Keywords: Enterprise Architecture, Bayesian Networks, Maturity Model, Czech e-Government, Digital Government.

Abstract: One of the current drivers for transitioning from the traditional E-Government to the digital government is

the ability to create and share new services in the governmental ICT landscape. The government must

effectively communicate and offer its services to itself (G2G) and outside, be it an end-consumer or business

(G2C, G2B). Since the government is internally divided, there is a need to measure its parts' performance for

effective management. However, conventional maturity models cannot address and explain the cause of the

differences, and thus typically respond to symptoms and show just winners and losers of the given benchmark.

From this position, a study and a deeper analysis of the maturity model used in the public administration of

Czechia are provided. Further analysis was undertaken via Bayesian networks to answer the question: How

do project management and prioritization affect service level management? Or how the enterprise architecture

as a method is linked to the overall organization's performance? Significant relationships were identified, and

the use of the Bayesian network as a prediction model was proposed. Further evaluation steps and research

opportunities were discussed.

1 INTRODUCTION

The main goal of the presented paper is to obtain and

compare the capabilities of individual public

authorities. Comparison is made from the viewpoint

of the National Architecture of the Czech Republic.

This overview of the current state of the scope of

capabilities of individual actors of public

administration is a suitable but also a necessary

starting point for the design of sustainable other

concepts, solutions, and development of national

architecture. Ultimately, national architecture is the

result of the cooperation of all individual actors.

Among the current problems of deploying and

maintaining corporate architecture in the public

sector are considered causes such as the resistance of

individual authorities to corporate architecture and

the division of roles, setting relevant goals, and the

issue of using corporate architecture in practice. A

partial problem pertinent to this area is the need for a

https://orcid.org/0000-0003-3336-265X

https://orcid.org/0000-0001-5810-4038

link between qualitative and quantitative data. Also,

specialized and expert experience and skills are

insufficient to manage complex systems (Seppänen et

al., 2018). Al-Kharusi et al. (2018), in a qualitative

case study of the public sector of Oman, elaborates

in-depth on the genesis of the creation of enterprise

architecture, in which the level of knowledge of

stakeholders and its sharing plays a key role.

2 THEORETICAL

BACKGROUNDS

Regarding the current state of maturity models, two

approaches predominate, or domain-specific meta-

models of maturity, for example, as Ostadzadeh and

Shams (2014) have shown in a study of highly

complex and interconnected systems for which

general meta-models were insufficient.

600

Rod, M. and Vomlel, J.

Improving Czech Digital Government Based on Quantiﬁed Maturity Model of Enterprise Architecture.

DOI: 10.5220/0011855300003467

In Proceedings of the 25th International Conference on Enter prise Information Systems (ICEIS 2023) - Volume 2, pages 600-607

ISBN: 978-989-758-648-4; ISSN: 2184-4992

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

As Santos-Neto and Costa (2019) introduced,

there is also an exciting trend where many models are

created, but only a part of them is validated and

applied.

2.1 Czech e-Government

From the point of view of the development of

enterprise architecture in the public administration of

the Czech Republic (Czechia), four years have passed

since the first significant standardization processes

commenced. These changes were driven by the

program Digital Czechia (Digitální Česko) (Czechia,

2018b) and strategically mainly with the Information

Strategy of the Czech Republic and its annexes

(Czechia, 2018a).

The annexes: National Architectural Framework

(NAF) and the National Architectural Plan (NAP),

are the backbone for the objectives of Digital

Czechia. This standardization follows the reforms and

concepts of managing the architecture of the public

administration as Hrabě (2013) presented in its work.

One of the central architectural governance

bodies is the Department of the E-Government Chief

Architect of the Ministry of the Interior of the Czech

Republic (shortened as DECA). DECA is also

responsible for approving ICT projects in the public

sector of Czechia.

2.1.1 Context of EU

The European Union has presented its framework for

measuring the maturity of service interoperability

across the members of the European Union (EU &

Mannot, 2016). The NAF acknowledges this

framework. However, the scope of the

interoperability framework is narrow, and usage is

limited. Currently, NAF includes the Benchmark of

Public Administration. However, that is a change

made in the past year. Up to that point, the NAF

included only brief subjective self-assessments based

on eight capabilities for managing an organization's

enterprise architecture, and the rest referenced

TOGAF (DECA, 2021).

As the TOGAF to this day does not contain its

maturity model, the reference was and still is a proxy

for the maturity models of the third parties (The Open

Group, 2019, 2022).

Another approach to measuring maturity in the

context of the European Union is The Digital

Economy and Society Index (DESI). DESI can be

used to ascertain Europe's overall digital performance

and the digital competitiveness of the corresponding

countries. DESI is a composite index of five

underlying categories (European Commission, 2022).

Looking at the relative position of countries,

Czechia is still underperforming and put behind the

average EU values. Moreover, Czechia is lacking in

digital public services.

Figure 1: DESI 2022 – and position of the Czechia

(European Commission, 2022).

2.2 Quantifying the Performance

The underlying problematization of given maturity

models lies in their explainability – with basic

questions such as: What is the cause? Why it works

that way? Conventional maturity models cannot

address and explain the reason for the differences.

Thus, typically respond to symptoms and show just

winners or losers of the given benchmark, where the

low-performing areas of the organization would be

targeted for improvement, but the real cause would

remain hidden.

From this position, the need for study and a

deeper analysis of the maturity model arises.

Explainability further raises the ambitions for internal

quantitative improvement based on additional

quantification and structuring of variables. This

approach is performed in the case study of the Czech

digital government (E-Government). This further

analysis was achieved via Bayesian networks,

theoretically described in the next chapter.

3 MATERIALS AND METHODS

This section introduces the primary data sources and

their processing methods. Firstly, the Benchmark of

the public administration of Czechia is described.

Then the approach of using Bayesian networks is

presented

3.1 Czech Benchmark of Public

Administration

The Benchmark of Public Administration, also known

as the ICT Benchmark of Public Administration, is

Improving Czech Digital Government Based on Quantiﬁed Maturity Model of Enterprise Architecture

601

Figure 2: Overall Maturity Level Dynamics – Change from 2018 to 2021. M(x) is an anonymized ministry, ÚSÚ(x) stands

for anonymized central administrative authority. Maturity could range from 1 to 5.

conducted every three years and is based on the

Digital Czechia program (Dzurilla et al., 2018). In

this paper, the Benchmark of the Public The

administration will be called "Benchmark". The

Benchmark focuses on three main areas: a) Public

administration management, b) Finance and

personnel situation, c) Subjective evaluation of the

Czech public administration and E-Government.

Each of these areas is broken down in more detail

into individual questions. The first Benchmark took

place towards the end of 2018. The second

Benchmark took place towards the end of 2021.

In terms of a scientific viewpoint, this Benchmark

is a domain-specific maturity model tailored for the

public sector, particularly the corresponding

legislation. This domain-specific approach is

justified, as Ostadzadeh & Shams (2014) showed in a

specific information system case study. In the case of

the complexity of public administrations and their

systems, the complexity will be higher. Of course,

this domain specificity is redeemed by a reduction in

interoperability. Thus, the possibility of ad-hoc

comparison of the different systems. It is important to

note that multiple domain-specific maturity models

are created; however, only a fraction is validated and

used (Santos-Neto & Costa, 2019).

The motivation for applying robust inference

techniques is to gain insight into the structure of the

issues or the concepts behind them. This could be

used to overview where knowledge needs to be added

or used correctly. Thus, it is a step forward in creating

and maintaining a functional enterprise architecture

approach (Al-Kharusi et al., 2018), but also the

possibility to create target points/states and find

scenarios that best support or even enable them.

An extensive benchmark was carried out from the end

of 2018. This Benchmark included all 14 ministries

of the Czech Republic and 20 central administrative

authorities (CAA for short).

In 2021, a second iteration of the Benchmark took

place where we, as the authors, were part of the

Benchmark team. Again, all the ministries were

surveyed. The overview of the Benchmark 2018 and

Benchmark 2021 can be seen below. In this paper, we

further explore the first and three main areas of

benchmarking: Level of governance, Level of change

management, and ICT level of governance and

capabilities.

Table 1: Overview of the used Benchmark 2018 and

Benchmark 2021 datasets.

Benchmar

2018 2021

Number of

anizations

14 ministries, 20

CAA

14 ministries, 20

CAA

(

7 rotated

)

Timeframe June to August

2018

October to

November 2021

Data

collection

Semi-structured

panel interview with

the

uestionnaire

Semi-structured

anel interview with

the

uestionnaire

Variables 40

(

as factors

)

(

as factors

)

3.2 Methods and Bayesian Networks

Firstly, the questionnaire data were evaluated using

descriptive statistics, namely sample means and

frequencies. As all forms were completed and no

gross errors were found, all the data were deemed

valid and were further used in the creation of the

Bayesian network.

ICEIS 2023 - 25th International Conference on Enterprise Information Systems

602

Table 2: Mapping the code questions/variables of the Benchmark 2021 and harmonized Benchmark 2018.

Code Question/variables

x0 <Dumm

variable> - ministr

(

)

, central administrative authorit

(

)

x1.1 Relevance and

ualit

of the or

anization's strate

(

existence and use of the Information Strate

gy)

x1.2.1 Up-to-date catalog of services and actions for citizens and companies (according to Law 12/2020 Coll.)

x1.2.2 Does a customer service manager exist to manage client services and service channels of the

ministr

/authorit

(

counters, e-filin

, data boxes, a

ortal of the ministr

/authorit

across a

as?

x1.3 Your or

anization's mana

ement s

stem, hierarchical vs.

rocedural mana

ement

x1.4.1 Degree and method of digitalization of agendas

x1.4.2 Degree and method of digitalization of support/operational processes (budgeting, human resources...)

x1.4.3 Degree and method of digitalization of management processes (planning, concept management, quality…

x1.5.1 Level of

ualit

/excellence mana

ement and feedbac

x1.5.2 Level of risk mana

ement

x2.1.1 Do you have a Digital Champion? (This is not a formal position of the Digital ambassador)

x2.2 ICT's position and mandate in the organization's management syste

x2.3.1 How is Enterprise Architecture (EA) maintained and used as a management method in the organization to

ort strate

lannin

chan

e mana

ement?

x2.3.2 Does the organization have its own internal Enterprise Architect?

x2.3.3 Are the new systems, or changes to systems, always approved by the unit of the Enterprise Architecture?

x2.4.1 How is project and program management used to deliver successful organizational change?

x2.4.2 Does the or

anization have

ect mana

ers?

x2.4.3 Is a

rocess for recordin

and

rioritizin

ects across the or

anization defined and used?

x2.4.4 Is there a dedicated (planned) capacity of systematized posts (or part-time posts) within the organization to

lement chan

(

for inclusion in

ects

)

x2.4.5 Is a process in place and routinely used to dedicate/release internal experts to projects and replace their

missing capacity in the line management of the organization's performance?

x3.1 Level of management of the information strateg

x3.2 Level of im

lementation of re

uirements mana

ement and its flow from the business to the ICT de

artments

x3.3 The current catalo

of internal IT services

x3.4 Service mana

ement in

lace

(

SLAs on all ke

stems, both to internal customers and external su

liers

)

x3.5.1 How have you addressed the integration of security policies into IT processes and procedures for the design,

lementation, o

eration, and use of IT solutions

x3.5.2 Have

ou im

lemented and activel

used SIEM

(

Securit

Information and Event Mana

ement

)

x3.6.1 IT

ualit

stem in

eneral

x3.7.1 Do you measure the cost per en

-user transaction?

x3.7.2 Do you measure the successful and unsuccessful completion of transactions?

x3.7.3 Do

ou measure user satisfaction with the a

lication/s

stem?

x3.7.4 Are

ou measurin

the usa

e of the di

ital channel versus the non-di

ital channel?

(

Where meanin

ful

)

x3.8 The abilit

of the IT de

artment to desi

n the s

stem, tender, and deliver on time with a

iven

ualit

x3.9 The ability of the Authority / IT Department to operate the systems and measure the quality of operation.

x3.10.1 Do you use the software provided as an external service (SaaS)?

x3.10.2 Do you use cloud solutions (running systems as a service, PaaS, IaaS)

x3.11.1 Do you have the source code for custom solutions and custom modifications to stand-alone software for all

critical IT solutions (primarily the legal category of public administration information systems)?

x3.11.2 Do you have development documentation (e.g., detailed data model) for all critical IT solutions?

x3.11.3 Do

ou have contractuall

secured licensin

hts to maintain and develo

IT solutions?

x3.11.4 Have you secured in-house competence (capacity and knowledge) to maintain and develop the organizations'

key platforms and solutions (for each solution accounting for at least 10% of the organizations’ IT spend)?

In addition to descriptive statistics, which

typically speak about the state, distribution, and

frequencies of the concepts under study, we wanted

to move within the knowledge modeling to a state that

would allow us to grasp the internal dependencies and

predict the conditions of individual ministries and

authorities. For this purpose, the approach of

Bayesian statistics and its Application using Bayesian

networks based on artificial intelligence and machine

learning was used.

The Bayesian network is a multidimensional

method that, in addition to the objectives defined

above, is user-friendly as part of its result is a

visualized graph (Jensen, 2001; Koller & Friedman,

2009). This graph can be imagined as a map of the

relationships between sub-constructs (variable), such

Improving Czech Digital Government Based on Quantiﬁed Maturity Model of Enterprise Architecture

603

as how a digital service uses the computing power of

the underlying servers and how these services relate

to the existing catalog of IT services. Simply put, the

network helps us find links between issues that would

not otherwise be visible.

The resulting Bayesian network can thus be seen

as a model that allows:

- Define the internal structure of the data.

- Analyse the probability distribution of the data.

- Predict data states of interest based on known

other related information.

The Bayesian network was used for the

underlying structure of the research concepts as a

learning algorithm has used a variety of score-based

(Hill-climb, tabu search) and constrain-based (pc-

stable, grow-shrink) suitable algorithms implemented

in the bnlearn (Scutari, 2018). For ascertaining, the

statistical significance chi-square test was used, and

for measuring the strength of the relationship, the

criterium of Mutual Information was used. Both those

metrics were based on the α = 5%, with a 95%

confidence interval. The Bayesian information

criterion (BIC) was used for the overall model quality

evaluation (Jensen, 2001; Koller & Friedman, 2009).

The algorithms were realized based on the bnlearn

package for R Scutari (2010).

4 RESULTS AND DISCUSSION

The input to the model to obtain the Bayesian network

was 34 vectors, all with 39 variables per vector. There

were no missing values. Given this number of

ministries and agencies surveyed, score-based gradient

algorithms performed well, whereby the best model

was then selected based on BIC. Final Bayesian

networks are computed via the Hill-Climb algorithm.

All relationships in the models are statistically

significant, where the mutual information parameter,

as was already mentioned, was used to classify and

determine the strength of the arc. The direction of the

arcs is defined based on internal network consistency

criteria. It thus cannot be considered as the direction

of causality, although it may be consistent with it. The

final model based on data from Benchmark 2021

contains three mutually disjoint Bayesian networks,

two of which are trivial, containing a maximum of

three elements. The last network includes a structure

of 21 nodes and 20 arcs. Due to the limited data set,

the created model is a tree (graph theory). We refer to

this Bayesian network as the “main network”.

No significant statistical relationships were found

between the ten questions (variables), so they are not

part of any of the networks mentioned in the model.

Let's analyze concrete questions such as question

x1.2.1 by looking at the answers, especially for the

ministries. We can see a possible pressure to answer,

corresponding to a clearly given legislative

obligation. This aspect of "not admitting weakness

and staying in the grey middle" needs to be

considered when assessing this Benchmark.

Figure 3: Model created from Benchmark 2021 data set.

A different perspective is offered when considering

question 2.3.1, “How is Enterprise Architecture (EA)

maintained and used as a management method…” as

this variable is independent and not connected to any

network.

This result indicates that the level of exercising

enterprise architecture does not bring a good effect,

as not even the dummy variable (x0) of the overall

maturity level is independent. This situation could be

explained as the enterprise architecture approach has

still failed to be adopted in today's public

administration of Czechia. The purpose of enterprise

architecture is the effective holistic management of an

organization.

Unfortunately, the current situation corresponds

to a situation where these thoughts are tightly linked

only to information systems. However, EA is not an

ICT discipline, although it is historically tied to it.

The situation where this approach is used from an ICT

direction is better than if it did not exist, but it is not

meant to be so. If this current ICT stigma were

removed and enterprise architecture departments had

access to ICT and non-ICT management and change,

the potential for effective functioning of single public

administration actors and their digital services would

be multiplied many times over.

4.1 Difference Between Ministries and

Central Administrative Authorities

Adding an auxiliary variable defining whether an

organization is a ministry or a CAA, we learn that

only two variables are affected by this division.

ICEIS 2023 - 25th International Conference on Enterprise Information Systems

604

The first of these behavioral differences is using

SIEM, one of the main tools for managing

cybersecurity and, thus, the reliable operation of

public administration. In terms of ministries, all but

one ministry used SIEM. The split in the central

administrative authorities is 60:40 in favor of using

SIEM. However, this situation cannot be considered

satisfactory; one of the priorities for the secure

functioning of the public administration would be to

roll out the already existing SIEM solutions to the

area of the UAS concerned and to introduce this

system unconditionally to the remaining ministries.

The second noticeable difference is the

representation of the measurement of the use of the

digital channel for services versus the non-digital

channel. Here, there are situations where the CCA

either excels (45%) or does virtually nothing (also

45%), with the remaining 10% (two CCA) of the

organizations surveyed falling somewhere in

between.

Looking at ministries, which are generally larger,

only one ministry indicates that it does so for all

meaningful activities. Most ministries (56%) need to

practice this measurement and evaluation. The

remainder, about a third of ministries, indicate that

measurement occurs when there is increased interest,

i.e., not routinely but systematically as needed.

However, measurement and evaluation are principles

of compiling and maintaining a catalog of digital

services, including planning new ones based on user

knowledge and thus facilitating the digitization of

such services.

4.2 The Second Bayesian Network

As another very trivial Bayesian network, the

relationship between the degree and manner of

digitalization of operational processes (question

1.4.2) and the measurement of user satisfaction with

the application/system is presented. In this condition,

a positive relationship is observed, where the

probability of measuring user satisfaction increases

with an increasing degree of digitalization of

operational processes.

The resulting model did not further connect this

network to other elements or directly to other

networks, but by looking deeper into the data

structure, a subjective connection, or hints of it, can

be found, at least with the other questions of the

service/transaction measurement capability topics

(3.7.1, 3.7.2). In the case of a more significant number

of data, this connection with other elements/networks

could be statistically confirmed based on the chosen

model criteria.

4.3 The Main Bayesian Network

The leading Bayesian network consists of 21

variables (maturity model questions). It can be

noticed that its visual structure corresponds to the

system of primary elements and the fans that branch

from them. Questions 2.4.4 examining how the

organization plans and allocates its internal capacities,

2.4.3 concerning the definition and prioritization of

projects, or 3.11.1 whether the organization has

solution source codes can be considered as the main

elements through which the remaining others are

linked.

Let us consider the last-mentioned element, i.e.,

the role of source codes in ICT solutions. The graph

shows that this element is statistically related to the

other four elements (questions 3.7.2, 2.4.1, 3.11.2,

3.11.3). If we mentally try to derive how the source

code solution will be related to the ownership of the

access documentation, it makes sense to have both

approaches at the same or similar level. It makes no

practical sense to have access to source code but no

longer to development documentation and vice versa.

Table 3: Conditional probability between the maturity of

questions 3.11.1 and 3.11.2.

3.11.1

maturity 1 3 5

3.11.2

1 0.75 0.07 0.00

3 0.25 0.87 0.40

5 0.00 0.07 0.60

A look at the probability ranking between the two

elements gives us the right idea. The most likely

situations are on the main diagonal (both maturities at

levels 1, 2, or 3). With that said, a relationship where

one maturity is at level 1 and the other at level 5 does

not occur.

The statement presented above could have been more

interesting. We had such an assumption beforehand,

so the result could be considered obvious. If we now

disregard the situation where this "obviousness" is

detrimental, we do get another, a more fundamental

piece of information. The model behaves as we

expect it to, i.e., the validity of the approach is

substantively demonstrated in this case. The

following sections will discuss only the conceptual,

possibly surprising implications.

Improving Czech Digital Government Based on Quantiﬁed Maturity Model of Enterprise Architecture

605

Figure 4: Bayesian Network Comparison Between Benchmark 2021 and Benchmark 2018.

4.3.1 Use of Software in the Form of SaaS

and Cloud Solutions

It is interesting that the two cloud questions (3.10.1,

3.10.2) are not together but are re-allocated to internal

management question 1.4.3 - The extent and manner

of digitalization of management processes..., and

question 2.4.4 - Is there a dedicated (planned)

capacity of systemized posts in the organization for

implementing change? Thus, under this assumption,

the way of using software such as SaaS and cloud

solutions are independent, given 1.4.3 or 2.2.4.

Those who can effectively plan and structure

internal capacity are able to use cloud solutions. The

reverse direction of the relationship is meaningless

here since the ability to delegate and compartmentalize

is independent of how the technology is implemented.

Conversely, of those who do not have this capability,

only one in two use cloud technologies. Thus, with the

progressive digitization of management pro-processes,

the ability to use the cloud can be influenced by

activating and enhancing the internal capacity planning

capabilities. A more in-depth analysis of the

management of the organizations in question would be

needed to determine what the specific steps should be.

However, in the first approach, the incremental

differences (deltas) between levels (maturity) can be

based on the answers obtained.

4.3.2 Prediction Capability

Suppose we are interested in how the level of the

service level management (3.4) is influenced by the

capability of the management of projects (2.4.3).

Figure 5: Main Bayesian Network – 2.4.3 Cut-Out.

As can be seen, there is a tight relationship. If the

capability is absent, the maturity level of 3.4 is more

spread and does not achieve the highest maturity. In

contrast, setting the evidence for 2.4.3 that the

organization's project management with prioritization

is exercised results in a higher chance for higher

maturity levels (Figure 5).

4.3.3 Comparison with the Benchmark 2018

For the comparison between 2021 and 2018, the

source data had to be harmonized first. Benchmark

2018 was harmonized to be comparable to

Benchmark 2021. The changes for Benchmark 2018

included removing or aggregating different questions.

By comparison, only two links have been

completely preserved (Figure 4, green relationships).

For the other nodes, there are changes (orange is

present in the Bayesian network from Benchmark

2021, and red arcs are present only in the Bayesian

network from Benchmark 2018). This difference can

be demonstrated by the already discussed usage of the

SIEM (question 3.5.2). Looking at the underlying

data of 2018, there was no difference between

ICEIS 2023 - 25th International Conference on Enterprise Information Systems

606

ministries and CAA. Ministries have acted, and

almost all of them have integrated the SIEM solution.

This overall significant difference between networks

calls for further exploration.

5 CONCLUSIONS

This paper addressed the quantification of

relationships within a standardized public

administration benchmark. Machine learning-based

Bayesian networks were used as a tool for this

quantification. Bayesian networks combine both

visual simplicity and explanatory and predictive

power. Also, the demonstrated approach is

generalizable.

By understanding the structure, strategic

decisions can be better directed, and processes of

digitalization and further development of the Czech

public administration can be made more efficient.

Further examination of the dataset and the

Bayesian network could bring more exciting findings

than those presented in this short paper. Also,

applying different approaches to aggregating the data

will enable different views on the matter. Moreover,

applying the leave-one-out cross-validation (Efron,

1982) for the presented model or constructing and

comparing more Bayesian network models could be

performed. A deeper evaluation of the differences and

their causes between Benchmark 2021 and

Benchmark 2018 could be another future topic.

Last but not least, insight could be gained with the

incorporation of the rest of the Benchmark available

data. The challenge would be making a hybrid

network with not just factor variables but also

numeric ones. As authors, we are excited about the

next Benchmark from the public administration of

Czechia and the possibility of further improving the

Czech Digital government.

ACKNOWLEDGEMENTS

This research was supported by the Internal Grant

Agency project of the Prague University of

Economics and Business IGS F6/61/2020. Also, it is

only proper to acknowledge the assistance of the

Department of the E-Government Chief Architect of

the Ministry of the Interior of the Czech Republic and

the partners from the program Digital Czechia and the

opportunity to use raw data from the Benchmark of

the Public Administration for the year 2018 and 2021.

REFERENCES

Al-Kharusi, H., Miskon, S., & Bahari, M. (2018). Enterprise

Architecture Development Approach in the Public

Sector: International Journal of Enterprise Information

Systems, 14(4), 124–141. https://doi.org/

10.4018/IJEIS.2018100109

Czechia. (2018) Usnesení Vlády České republiky č. 629 k

programu Digitální Česko a návrhu změn Statutu Rady

vlády pro informační společnost.

Czechia. (2018). Informační koncepce České republiky.

https://www.mvcr.cz/soubor/vladni-program-

digitalizace-ceske-republiky-2018-digitalni-cesko-

informacni-koncepce-cr.aspx.

Efron, B. (1982). The Jackknife, the Bootstrap and Other

Resampling Plans. Society for Industrial and Applied

Mathematics. https://doi.org/10.1137/1.9781611970319

EU, & Mannot, P. (2016, March 27). Interoperability

Maturity Model | Joinup. https://joinup.ec.europa.eu/

collection/semantic-interoperability-community-

semic/document/interoperability-maturity-model

European Commission. (2022, July 28). Digital Economy

and Society Index (DESI) 2022 | Shaping Europe’s

digital future. https://digital-strategy.ec.europa.eu/en/

library/digital-economy-and-society-index-desi-2022

Hrabě, P. (2013). Koncepce podnikové architektury pro

reformu veřejné správy ČR [Doctoral, VŠE].

https://vskp.vse.cz/47869_koncepce_podnikove_archite

ktury_pro_reformu_verejne_spravy_cr

Jensen, F. V. (2001). Bayesian networks and decision graphs.

Springer.

Koller, D., & Friedman, N. (2009). Probabilistic graphical

models: Principles and techniques. MIT Press.

DECA (2021). Národní architektonický rámec [Architektura

eGovernmentu ČR]. https://archi.gov.cz/ nar_dokument

Ostadzadeh, S., & Shams, F. (2014). Towards a Software

Architecture Maturity Model for Improving Ultra-Large-

Scale Systems Interoperability. ArXiv: 1401.5752 [Cs].

https://doi.org/10.7321/jscse.v3.n3.13

Santos-Neto, J. B. S. dos, & Costa, A. P. C. S. (2019).

Enterprise maturity models: A systematic literature

review. Enterprise Information Systems, 13(5), 719–769.

https://doi.org/10.1080/17517575.2019.1575986

Scutari, M. (2010). Learning Bayesian Networks with the

bnlearn R Package. Journal of Statistical Software, 35(3).

https://doi.org/10.18637/jss.v035.i03

Scutari, M. (2018, November 8). bnlearn—Man/bnlearn-

package.html. http://www.bnlearn.com/documenta

tion/man/bnlearn-package.html

Seppänen, V., Penttinen, K., & Pulkkinen, M. (2018). Key

Issues in Enterprise Architecture Adoption in the Public

Sector. Electronic Journal of E-Government, 16(1), 46–

58.

The Open Group. (2019). The TOGAF Standard, Version

9.2—Architecture Maturity Models. https://pubs.open

group.org/architecture/togaf9-doc/arch/chap45.html

The Open Group. (2022). TOGAF Series Guide: Maturity

Models. https://pubs.opengroup.org/togaf-standard/

architecture-maturity-models/

Improving Czech Digital Government Based on Quantiﬁed Maturity Model of Enterprise Architecture

607