Searching for Weak Signals in the Web to Support Scenarios

Building for Future Studies

Rodrigo Teixeira dos Santos

, Edilson Ferneda

, Hercules Antonio do Prado

Aluizio Haendchen Filho

, Ana Paula Bernardi da Silva

, Roseane Salvio

and

Elaine Coutinho Marcial

Catholic University of Brasilia, Brasilia-DF, Brazil

University of the Itajai Valley, Itajai-SC, Brazil

Federal Institute of Brasília, Brasilia-DF, Brazil

Mackenzie Presbyterian College of Brasilia, Brasilia-DF, Brazil

Keywords: Future Studies, Foresight, Scenarios, Weak Signals, Multi-agent Systems, Early Warning Systems.

Abstract: A specification of a multi-agent system for searching weak signals on Internet is proposed for supporting

future studies. A set of software artifacts are presented, including: (i) functional/non-functional requirements; (ii)

use-cases involving human and software agents; (iii) a database model containing the main entities; and (iv)

a role and functional model with the main interactions and activities the agents are involved. A prototype were

developed and evaluated by experts in future studies. From this specification/prototype, an early warning

system can be developed to support intelligence analysts in producing qualified information for future studies.

1 INTRODUCTION

Concerns on future have always been on human race

focus. It is probably more intense in the current days

due to the accelerated rhythm of change and the

increasing uncertainty as consequence of the

downfall of the well-regulated world (Jouvenel,

2000). Along the ages, the fear had been an important

mechanism for mind changing of the human being,

leading to the acknowledgment that the future is not

pre-determined or written, but rather can be built by

the actions of social actors. In this way, people had to

learn how to live under uncertainties, observing that

the future emerges from seeds (or signals) left in the

past or present (Marcial and Grumbach, 2002). These

seeds must be captured and assessed for building

knowledge about the future as done in the disciplines

Strategic Prospective and Competitive Intelligence

(CI).

https://orcid.org/0000-0002-2138-8126

https://orcid.org/0000-0003-4164-5828

https://orcid.org/0000-0002-8375-0899

https://orcid.org/0000-0002-7998-8474

https://orcid.org/0000-0002-9963-282X

https://orcid.org/0000-0001-9172-8299

https://orcid.org/0000-0001-9686-8418

Scenarios Building is a tool for studying the

future. It helps to give a long-term approach in a

world of uncertainties (Schwartz, 1991). Moreover, it

stimulates strategic thinking and helps to run over

thinking limitations by means of multiple futures

envisioning (Ammer, Daim, and Jetter, 2013).

Godet and Roubelat (1996) argues that “multiple

and several potential futures are possible; the path

leading to this or that future is not necessarily unique.

The description of a potential future and of the

progression towards it comprises a scenario”.

For Marcial and Grumbach (2002), Scenarios

Building is an important tool for generating and

assessing strategic definitions in the increasingly

turbulent and uncertain environment we live in.

Moreover, it facilitates the communication

concerning visions about the future, unifying

organizational language, helping the creativity

development, and the organizational learning. It also

Santos, R., Ferneda, E., Antonio do Prado, H., Filho, A., Bernardi da Silva, A., Salvio, R. and Marcial, E.

Searching for Weak Signals in the Web to Suppor t Scenarios Building for Future Studies.

DOI: 10.5220/0010481609010908

In Proceedings of the 23rd International Conference on Enterprise Information Systems (ICEIS 2021) - Volume 1, pages 901-908

ISBN: 978-989-758-509-8

901

improves the understanding on the organizational

surroundings, allowing to envision the possible risk

situations and uncertainties.

Scenarios Building can be considered under two

complimentary approaches (Godet, 1986):

(i) exploratory, the ones that highlights actual or past

trends and that may draw a possible future; and (ii)

normative/anticipative, that takes the opposite route,

starting from the future to the present, thereby

inducing the factors of evolution.

An important kind of information for scenarios

planning is the weak signals, defined by Ansoff

(1982) as “external or internal warnings that are too

incomplete to allow in precision their impacts or an

estimate response to them”.

The connection between the Scenarios Building

Method and the analysis of weak signals can occur in

the context of an intelligence activity. It involves

production of information integrated, assessed, with

high aggregate value, and in the most part, carrying

seeds of future (Marcial and Grumbach, 2002). For

short, Scenarios Building Method relates to CI in the

sense that the former produces future views and the

path that leads to them, and the latter helps with the

production of pieces of possible information.

Additionally, Prescott (1995) states that CI is

more than a product, it is a process that helps the

organization in the anticipation of threats and

opportunities, in order to take decisions in a securer

way and make your competitive environment better

monitored. The use of CI is endorsed by the

aggregated business value in the endeavor, pursuing

a competitive advantage. CI is a legal and ethical

process of collection and informational evaluation

regarding the competitors and industries in operation

to help an organization in taking better decisions and

reach its goals (Bergeron and Hiller, 2002).

Marcial and Grumbach (2002) recognize the

connection between CI with scenarios building and

strongly recommends its application to keep tracking

of actors or variables that might be important to

achieve organizational objectives.

The widespread of mobile devices and social

networks in the last years enabled the creation and

transmission of digital contents in astonishing levels.

This huge amount of information has made hard the

companies to identify weak signals. In this paper, an

approach for identifying and monitoring this kind of

information in Internet is proposed.

2 BACKGROUND

2.1 Weak Signals

A signal is an event in which any actor in the

environment sends a message while executing an

activity or as a result of a specific action (Coffman,

1997). When a message is sent, someone or

something may receive and understand it. There are

three types of signal (Coffman, 1997): (i) the ones

beyond perception; (ii) the perceptible, but unknown

due of the receptor’s mindset; (iii) those that are

known by the receiver’s and are used for a behavior

change. The first ones are just signals that are

continuously sent by the source, but not received. The

second ones are perceptible by the receiver but for

some reason they are ignored. The third signals type

are those signals that are perceived by the receiver

and is used to adjust its own course, actions or

behaviors.

In a philosophical sense, Jouvenel (2000) says that

“the future leaves in the past and in the present, seeds

that might germinate or not […] transforming into

great fruit trees, plants that will never bear fruits or

weeds”. One of these seeds is those facts that carriers

the future and are characterized by being apparently

unimportant, but with relevant consequences and

potentialities. These are the weak signals.

One of the main objectives when dealing with

weak signals, is to take awareness of the organization

surrounding, allowing it to keep an advantageous

position with respect to its competitors (Janissek-

Muniz and Blanck, 2014). Furthermore, these

situations lead to perceptions of possible future

situations that might be interesting for the

organizational evolution (Marcial and Grumbach,

2002). Being aware of these signals may drive the

organization to possible different futures like an

innovation or a trend that affects business and its

environment (Coffman, 1997). It includes threats or

opportunities, learning lessons, growing, and

developing.

In order to search for weak signals for envisioning

the future, it is important to consider that these pieces

of information are "anticipative, qualitative,

ambiguous, fragmented and may come in various

formats and from distinct sources” (Fonseca and

Barreto, 2012). Additionally, Marcial and Grumbach

(2002) emphasize the importance of perceiving the

mindset involved in the signal signification process.

A process for searching anticipative signals were

proposed by Fonseca and Barreto (2012) and

comprises the following steps: (i) stimulus

perception; (ii) interpretation for sense making of the

ICEIS 2021 - 23rd International Conference on Enterprise Information Systems

902

identified signals; (iii) learning or incorporation of a

new information in a company database. The

importance of this process for an organization is

highlighted by Freitas and Janissek-Muniz (2006) for

generating warnings when some key intelligent topic

is modified. During the process, some questions must

be answered: What information has been searched?

What type of information has been searched? Whom

is concerned about the information? What is the

information source? How it was obtained? Why is the

information important? What is the objective of the

information?

With respect to the sources of weak signals,

Marcial and Grumbach (2002) considers two

possibilities: formal sources, that involve books,

reports, laws, normative papers, patents, Internet,

among others, and the informal, that include

interactions with competitors, clients or providers,

and participation in congress or seminars.

2.2 Early Warning Systems and Search

on Internet

The interest on Early Warning Systems (EWS)

emerged in the sixties as an alternative for

international policies change predictions (Rausch et

al., 2012). Later, in the seventies, the ability of

offering a timely response to a problem began to be

considered as an important characteristic of EWS

(Ansoff, 1980).

Ansoff (1980) argues that organizations have

always faced a paradox when dealing with issues

apparently unimportant: usually they wait until the

fact is better known to adapt their strategic planning.

Ansoff and McDonnell (1990) proposed the Issue

Strategic Management System that deals with

unexpected change in three ways: (i) Detecting

surprises in real time; (ii) Responding in the moment

that they occur without updating the strategic

planning; and (iii) Creating a problem solving cross

organizational task force.

Dohn et al. (2013) suggest the following

requirements when considering the Internet as a

context for EWS searching: (i) To have a high amount

of storage and processing capacity in order to make

possible the integration of potentials weak signals; (ii)

Being flexible enough to make possible changes; (iii)

Having integrating tools with quantitative and

qualitative characteristics; (iv) Enable the

information sharing with stakeholders interested in

the same issue; (v) To have usability and portability;

(vi) To enable the users integration; (vii) To define

actions from weak signals; and (viii) Stimulating the

organizational learning.

Due to the textual nature of most information in

the Internet, the search for weak signals on there can

benefit from text analysis techniques like Text

Mining and Natural Language Processing. Mainly

considering that the Internet consumes too much time

from the users when they need to find some

information (Dkaki, Dousset, and Mothe, 1997).

Weiss, Indurkhya, Zhang (2010), Meystre et al.

(2008) reinforce the advantages of these techniques

for solving problems like documental classification or

organization, information retrieval or extraction,

clustering analysis, assessment, prediction, and

named entity recognition.

Other important tool to be considered for

searching weak signals in the Internet are the web

crawlers. Web crawlers are a kind of robot that cross

the web searching for information that might be

indexed and retrieved (Kumar, Bhatia, and Rattan,

2017). Figure 1 shows a generic architecture for a

web crawler. Web crawling starts from an initial URL

and seeks for other links aiming to retrieve the most

webpages and web documents related to a topic.

Figure 1: Generic architecture of a web crawler (Kumar,

Bhatia, and Rattan, 2017).

3 RELATED WORKS

The literature regarding EWS techniques is related to

issues like CI, corporate foresight, decision making,

early warning, early detection, environmental

scanning, foresight, horizon scanning, scenario,

social network, and strategic planning.

Saritas and Burmaoglu (2015) point out that, in

the EWS context, the amount of methods, both

quantitative and qualitative, is increasing, including

the ones applied to foresight practices. In the same

sense, Muhlroth and Grottke (2018) notices a

strengthening in the research related to EWS and

emphasize the importance of investments in

improving search engines, releasing human effort for

more advanced stages of studies.

Searching for Weak Signals in the Web to Support Scenarios Building for Future Studies

903

Schuh et al. (2016) proposed a generic EWS

model, emphasizing that these systems must produce

results that allow an organization to take specific

actions and learn from them.

Steinecke, Quick, and Mohr (2011) suggested six

points to be considered when implementing a EWS

system: (i) visualizing both, the internal and external,

organization environment; (ii) the existence of

indicators to monitor causes and effects; (iii)

automatize the indicators management; (iv) proceed

to a benchmarking of implementations; (v) to apply

the results found.

Rohrbeck, Thom, and Arnold (2015) argues that

an important requirement for a EWS is the ability to

integrate information from different people, enable

the emergence of knowledge from the confront

among ideas and connect this result to a specific

organizational process.

For Gheorghiu, Andreescu, and Curaj (2016), an

EWS must keep good early signals sources, efficient

information filters and a consistent characterization

of those signals.

A great variety of techniques for EWS can be

found. Garcia-Nunes and Silva (2018) aggregates

ontologies for text analysis in order to provide

semantic analysis. Kim, Park, and Lee (2016) apply

keyword-based analysis for envisioning weak signals.

Dousset, Elhaddadi, and Mothe (2011) emphasizes

that weak signal orbits near these keywords. Semantic

Analysis is also explored by Griol-Barres, Milla, and

Millet (2019), combined with quantitative textual

analysis, as cooccurrence of terms. This technique is

also applied by Kim et al. (2019). Along with web

crawlers, text analysis tools are applied for searching

on the Internet (Garcia-Nunes and Silva, 2018; Kim

et al., 2019).

Schuh et al. (2016) propose a generic model for

EWS, emphasizing that such systems must provide

qualified information for decision making support,

beyond strengthening the organizational learning.

Steinecke, Quick, and Mohr (2011) identified

some issues to be considered when implementing an

EWS: (i) watching internal and external

environments; (ii) availability of causal indicators;

(iii) apply Information Technology for managing

indicators and the information sources to be

monitored; (iv) search for complimentary and

contextual information for supporting results

analysis; (v) managing the results; and (vi) create

procedures able to update systems and share the

findings.

Rausch et al. (2012) recommended that this kind

of system must comply with the cultural

characteristics of each country.

For Dohn et al. (2013), an EWS can support an

organization by: (i) integrating many methods and

concepts for detecting early signals; (ii) information

management; (iii) causal networks interpretation.

In their analysis of an EWS, Rohrbeck, Thom, and

Arnold (2015) consider relevant the system cope with

information from many people, issuing knowledge

from different points of view and enabling the

connection of this knowledge with an organizational

process.

Gheorghiu, Andreescu, and Curaj, 2016) adds

good filtering processes and a categorization of weak

signals as requirements for a successful EWS.

4 ANALYSIS AND DESIGN

The previously discussed approaches focus on

searching on the Internet for enforcing the soundness

of a hypothesis. The proposal here presented is

focused in prospection of signals with no prior

definition of an inquire but inside a defined interest

domain. The proposal design benefits from the low-

coupling and flexibility of multi-agent systems

(MAS). Under a functional point of view, the choice

for MAS technology consider the possibility of

having some level of intelligence on the search

process. The analysis phase comprises the choose of

the development tool and the system requirements

specification. The project phase consists in define the

functionalities and the generic architecture of the

system and the data model.

4.1 Adopted Platform

It was adopted the MIDAS (Haendchen Filho, 2021)

framework for the implementation of the prototype.

The architecture is based on the coexistence of

several containers that communicate by means of a

front-end server. Each container provides an

environment for development and execution of

agents. Figure 2 shows the platform generic

architecture.

The framework consists of three layers: (i)

Middleware, represented by the entities Broker,

Proxy, Catalog, Manager, and Blackboard; (ii)

Services; and (iii) Agents. The middleware layer

facilitates the development by abstracting complex

developer procedures. It provides communication,

concurrency, lifecycle management, and discovery.

The Agents and Services layers enables development

of microservices agent-based applications.

ICEIS 2021 - 23rd International Conference on Enterprise Information Systems

904

Figure 2: MIDAS Architecture.

The applications are instantiated from the Agent

Container (AC

… AC

). The communication

between Agent and Service layer is done by Proxy. An

agent does not need to know who the service provider

is. This feature ensures transparency and decoupling

from the implementation. Services can be

independently created, and agents can use and

coordinate these services in their workflow.

The platform uses three interfaces: (i) HTTP for

inter-platform communication among the front-end

server AS and the AC containers; (ii) API interface,

which enables the communication with external

applications via REST protocol; and (iii) a WEB-

Interface for platform management and configuration

by developers and stakeholders.

4.2 Functional Requirements

Nine functional requirements were considered for

implementing the solution: (i) Identifying domains for

categorizing weak signals; (ii) Automatic search for

crossing the web for finding information (crawlers);

(iii) Keeping pages contents for later analysis; (iv)

Results analysis, in order to elaborate knowledge

from data stored; (v) Storing final results, necessary

for persistency of a signal; (vi) Websites list,

important for defining a start point for a search; (vii)

Storage and structure, necessary for keeping

complimentary information; (viii) Integration, to

enable interchange of information among the people

involved; and (ix) Access to the tool, including

usability and ability to run on multiplatform. The

human actors are: (i) Standard User, (ii) Manager of

Studies, and (iii) Expert. The software agents,

instantiated from MIDAS, are: (i) AgLINK, (ii)

AgWEB, (iii) AgRESULT, and (iv) AgWORD.

The most important use cases extracted from the

relation among requirements and respective actors

are: (i) Maintain search domain (Manager of

Studies); (ii) Visualize my studies (Manager of

Studies); (iii) Visualize the ongoing studies (Manager

of Studies); (iv) Login the tool (Standard User); (v)

(Manager of Studies); (vii) Maintain stopwords

(Manager of Studies / Expert); (viii) Maintain

dictionary

(Manager of Studies / Expert); (ix) Invite

for discussion (Manager of Studies); (x) Maintain

discussion (Manager of Studies / Expert); (xi)

Manage signal (Manager of Studies / Expert); (xii)

Generate word maps (AgWORD); (xiii) Maintain

preliminary results (AgRESULT); (xiv) Maintain

Web information (AgLINK / AgWEB).

4.3 Generic Architecture

Figure 3 shows the main elements of the architectural

Figure 3: System generic architecture.

Searching for Weak Signals in the Web to Support Scenarios Building for Future Studies

905

model, including the inter-relations among actors.

Standard User represents people involved in any

study, Manager of Studies is responsible for keeping

track of information on the studies, and Expert is any

domain expert invited to contribute to a specific

study. AgLINK is responsible for searching and

inserting new links in the database, AgWEB recovers

information from the registered web links,

AgRESULT, matches the pages content with the

query, beyond executing data cleaning, and AgWORD

is responsible for verifying the cooccurrence of words

in the documents and creating the word maps.

The process starts with the login of the Study

Manager. Following, two non-exclusive paths are

possible: (i) in the case of previous existing studies,

with pending searches, related to the same manager,

AgLINK activates AgWEB to retrieve and maintain

the related information, and (ii) the manager registers

a study, including the definition of the focus with the

respective links, allowing AgLINK to notify AgWEB

to access the links and keep the information retrieved.

This agent checks for the existence of links recorded

in the database. Next, the page content is inserted in

the database by request of AgWEB, that notifies

AgLINK and AgRESULT to access the webpage

content and search its weblinks. AgRESULT takes the

inserted content and proceed to preprocessing actions.

After that, AgWEB interactions are completed.

4.4 Data Model

The application handles 15 tables accounts of the data

model required for persistence: (i) the study domain,

(ii) status of the study, (iii) dictionary for keeping

synonyms; (iv) stopwords; (v) website links; (vi)

relationships among the studies and links;

(vii) studies; (viii) words related to the studies;

(ix) words related to a possible signal; (x) weak

signals; (xi) relationships among words and the

webpages; (xii) recovered links in a webpage;

(xiii) interchange of messages among managers of

studies and invited experts; (xiv) invited experts; and

(xv) contributions of invited experts in a discussion.

Figure 4 shows de tables and its relationships and

main attributes.

5 PROTOTIPATION

A proof of concept was developed under MIDAS in

order to evaluate the specification. The development

process includeds four steps described next.

I. Parameterization of Internet Search. The first

step is to insert information to drive the search. A

search expression must be provided to works as a

guess for finding information in the Web. The search

may be linked to a domain (technological,

governmental, environmental, etc.). Optionally, some

sites may be informed to be accessed as source of

information.

II. Retrieving Information from the Internet. After

the search begins, agents take action by scanning the

database. AgLINK accesses each of the links found in

Figure 4: Data model and relationships.

ICEIS 2021 - 23rd International Conference on Enterprise Information Systems

906

the database in order to find more links and register

them. At the same time, AgWEB retrieves the links

and accesses the content, storing them in a database.

III. Analyzing the Recovered Information. As the

website information is collected and stored,

AgRESULT starts the pre-processing. This involves

differentiating content and html tags, removing

stopwords, unifying words or synonyms, removing

scores, and checking whether the content is linked to

the search. After preprocessing, a quantitative

analysis is carried out by AgWORD, that produces a

word map (Figure 5). It is the main analysis tool

issued to the participants of the study.

Figure 5: Word map GUI.

IV. Using the Prototype. The participants were

invited to evaluate the tool by navigating in it

functionalities, exploring creation of word maps and

the signals that may arise from this representation.

The identification of a weak signal is completed after

a discussion that takes into account the different

views on the word map.

Figure 6: Inserting a weak signal.

A weak signal is registered as shown in Figure

6. One of the basis for getting insights from a weak

signal is to establish an active discussion, under

different perspectives and with different actors

(Mendonça, Cardoso and Caraça (2012).

6 FINAL REMARKS

This paper faces the challenge of identifying weak

signals with potential to be disruptive in a possible

future and a certain domain. The huge and increasing

volume of information the intelligence activity has to

deal with in this context can be widely facilitated by

this kind of computational support.

It is proposed a MAS conceptual model and a

design, including requirements, actors, and use cases,

as a solution for this problem. A prototype was

developed as a proof of concept and an evaluation by

experts in future studies was carried out.

Some future works for enriching such

specification are (i) the inclusion of other sources of

information (e.g., social networks), (ii) adoption of an

ontology in order to have a uniform vocabularies for

each domain.

REFERENCES

Ammer, M., Daim, T., Jetter, A., 2013. A review of

scenario planning. Futures, 46: 23-40.

Ansoff, H. I., 1980. Strategic issue management. Strategic

Management Journal, 1(2), 131-148.

Ansoff, H. I., 1982. Strategic response in turbulent

environments. Brussels: European Institute for

Advanced Studies in Management.

Ansoff, H. I., McDonnell, E., 1990. Implanting strategic

management. Prentice Hall.

Bergeron, P., Hiller, C. A., 2002. Competitive intelligence.

Annual Review of Information Science and Technology,

36(1), 353-390.

Coffman, B. S., 1997. Weak signal research - Part II -

Information theory. http://legacy.mgtaylor.com/

mgtaylor/jotm/winter97/infotheory.htm

Dkaki, T., Dousset, B., Mothe, J., 1997. Mining information

in order to extract hidden and strategic information.

RIAO'97: Computer-Assisted Information Searching on

Internet, pp. 32-51.

Dohn, K., Gumiński, A., Matusek, M., Zoleński, W., 2013.

The early warning concept in the system supporting

knowledge management in Polish machine-building

industry enterprises. KIM 2013, Knowledge and

Information Management Conference: Sustainable

Quality, pp. 211-221.

Dousset, B., Elhaddadi, A., Mothe, J., 2011. Content

accessibility and semantic networks processed on

foreign natural language analysis. Journal of

Intelligence Studies in Business, 1(1), 5-18.

Searching for Weak Signals in the Web to Support Scenarios Building for Future Studies

907

Fonseca, F., Barreto, L., 2012. Processo decisório e o

tratamento de sinais fracos. Future Studies Research

Journal: Trends and Strategies, 3(2), 34-58.

Freitas, H., Janissek-Muniz, R., 2006. Proposal of a

platform for competitive intelligence. Congresso Ibero-

Americano de Gestão do Conhecimento e Inteligência -

GeCIC, I. Curitiba, Brazil. (in Portuguese)

Garcia-Nunes, P., Silva, A., 2018. Using a conceptual

system for weak signals classification to detect threats

and opportunities from Web. Futures, 107, 1-16.

Gheorghiu, R., Andreescu, L., Curaj, A., 2016. A foresight

toolkit for smart specialization and entrepreneurial

discovery. Futures, 80, 33-44.

Godet, M., 1986. Introduction to la prospective: seven key

ideas and one scenario method. Futures, 18(2), 134-

157.

Godet, M., Roubelat, F., 1996. Creating the future: the use

and misuse of scenarios. Long Range Planning, 29(2),

164-171.

Griol-Barres, I., Milla, S., Millet, J., 2019. Improving

strategic decision making by the detection of weak

signals in heterogeneous documents by text mining

techniques. AI Communications, 32(2), 1-14.

Haendchen Filho, A., 2021. A Lightweight Microservice-

Oriented Platform for Development of Intelligent

Agent-based Enterprise Applications. ICAART:

International Conference on Agents and Artificial

Intelligence (to appear)

Janissek-Muniz, R., Blanck, M., 2014. Weak signals

management, entrepreneurship and uncertainty: a

relational theoretical essay under the perspective of

intelligence. 11

CONTECSI International Conference

on Information Systems and Technology Management.

Jouvenel, D. H., 2000. A brief methodological guide to

scenario building. Technological Forecasting & Social

Change, 65(1), 37-48.

Kim, H., Han, Y., Song, J., Song, T. M., 2019. Application

of Social Big Data to identify trends of school bullying

forms in South Korea. International Journal of

Environmental Research and Public Health, 16(14).

Kim, J., Park, Y., Lee, Y., 2016. A visual scanning of

potential disruptive signals for technology

roadmapping: investigating keyword cluster, intensity,

and relationship in futuristic data. Technology Analysis

& Strategic Management, 28(10), 1-22.

Kumar, M., Bhatia, R., Rattan, D. (2017). A survey of web

crawlers for information retrieval. Data Mining and

Knowledge Discovery, 7(6).

Marcial, E. C., Grumbach, R. J. S., 2002. Prospective

scenarios: how to build a better future. FGV Editora.

(in Portuguese)

Mendonça, S., Cardoso, G., Caraça, J., 2012. The strategic

strength of weak signal analysis. Futures, 44(3), 218-

228.

Meystre, S. M., Savova, G., Kipper-Schuler, K., Hurdle, J.,

2008. Extracting information from textual documents in

the electronic health record: a review of recent research.

Yearbook of medical Informatics, pp. 128-144.

Muhlroth, C., Grottke, M., 2018. A systematic literature

review of mining weak signals and trends for corporate

foresight. Journal of Business Economics, 88, 643-687.

Prescott, J. E., 1995. The evolution of Competitive

Intelligence. International Review of Strategic

Management, 6, 71-90.

Rausch, A., Fafaliou, I., Bedenik, N. O., Labas, D., 2012.

Early warning systems - empirical evidence. Review of

Marketing Theory and Practice, 24(2), 201-218.

Rohrbeck, R., Thom, N., Arnold, H., 2015. IT Tools for

foresight: the integrated insight and response system of

Deutsche Telekom innovation laboratories.

Technological Forecasting and Social Change, 97,

115-126.

Saritas, O., Burmaoglu, S., 2015. The evolution of the use

of Foresight methods: a scientometric analysis of global

FTA research output. Scientometrics, 105(1).

Schuh, G., Kabasci, P., Drescher, T., Ryschka, S.,

Wetterney, T., 2016. Identifying and evaluating

disruptive technologies using technology scanning.

International Association for Management of

Technology (IAMOT Conference).

Schwartz, P., 1991. The art of the long view. Broadway

Business.

Steinecke, N., Quick, R., Mohr, T., 2011. Environmental

scanning systems: state of the art and first instantiation.

Pacific Asia Conference on Information Systems,

PACIS 2011: Quality Research.

Weiss, S., Indurkhya, N., Zhang, T., 2010.

Fundamentals

of Predictive Text Mining. London: Springer.

ICEIS 2021 - 23rd International Conference on Enterprise Information Systems

908