Chapter 4

Collaborative Information Seeking in the Competitive Intelligence Process¹

4.1. Introduction

In a world undergoing perpetual social and economic transformation, decision making demands permanent reactivity, particularly in light of the globalization phenomenon. Decision making is the core of competitive intelligence (CI) and requires us to seek information that could help in resolving a decision problem (DP). In most books on decision support, emphasis is placed on the importance of clarifying a DP and on information seeking (or gathering), which are the primordial phases in solving a DP.

Going from the phase of clarification of the DP to the information seeking phase requires that the DP be first transformed into an information problem (IP). Thus, an IP can be considered to be a derivative of a DP [KIS 07, ODU 09].

An IP demands that we seek and retrieve information, indispensable activities in solving a DP following the CI process. However, it has been shown that the final goal of information retrieval is the production of knowledge, and that this is a collaborative task¹ [KAR 98]. Several studies have shown that information retrieval is both a cognitive and a social process that requires collaboration between users. In fact, users’ information behavior reveals that they manifest collaborative information seeking (CIS) patterns when attempting to solve IPs.

Based on the fact that the overall effectiveness of a group is not the same as the sum of its individual parts, we surmise that IPs, particularly in the context of CI, may be better and more rapidly resolved if a CIS approach is used.

We consider that, whether in designing methods or developing tools, information seeking and retrieval should not be seen as an individual activity but rather as a collaborative act. We will therefore attempt, in this chapter, to demonstrate the importance of CIS, on the one hand, and, on the other, a way of managing CIS in solving IPs, particularly in today’s world where collaborative watch is beginning to gain ground in socioeconomic sectors.

To explain what we mean by CIS and how this may be applied to the CI process, we will begin by presenting CI processes as seen by the SITE research team at LORIA laboratory. Second, we will present the evolution of information retrieval toward collaborative information retrieval (CIR). Based on the information behaviors of users, we will consider the CIS, and retrieval as a social and cognitive activity that leads to knowledge sharing. The third section will focus on approaches for facilitating and managing CIS.

4.2. The CI process

CI can be defined as the set of coordinated actions of seeking, processing, and disseminating useful information to economic actors susceptible to use this information. These actions are carried out legally, with all necessary guarantees for the preservation of business patrimony, in the best possible quality, time, and cost conditions [MAR 94]. According to Revelli [REV 98], CI is the process of collecting, processing, and disseminating information with the goal of reducing uncertainty in any strategic decision making process.

CI is linked to various similar concepts, such as economic intelligence (which is the term commonly used in French), business intelligence, and knowledge management. It is a process that embodies decision making. Decision making can be regarded as an outcome of mental processes (cognitive processes) leading to the selection of a course of action among several alternatives [THI 02]. CI is also considered to be an information process, made up of the following phases [DAV 09]:

a) identification of a DP;

b) transformation of the DP into an IP;

c) identification of relevant sources of information;

d) collection of relevant information;

e) analysis of collected information to extract indicators for decision making;

f) interpretation of indicators;

g) decision making.

Our focus here is not to elaborate on the CI process but rather to highlight the level of involvement of CIS in the CI process. We are interested in phases b, c, d, e, and f which are the phases in which the CIS activities are carried out by CI actors. However, it is important to note that there are four invariable factors in CI processes: the decision maker, the DP, information, and the protection of material and immaterial patrimony [DAV 09]. In a CI project, three main actors cooperate to ensure the success of the process: the decision maker, the watcher, and the coordinator.

The decision maker, who sits at the top of the process, is capable of identifying the DP to be solved in terms of stakes, risks, or threats to the company. The watcher, on the other hand, is responsible for the collection, analysis, and dissemination of information throughout the company. The coordinator acts as an intermediary between all CI actors and is responsible for managing workflow in the CI process, and for coordinating interactions among actors [KNA 07].

Philippe Kislin, in his PhD thesis [KIS 07], analyzed the cooperation process that may likely be established between a decision maker and a watcher in transforming a DP into an IP. Indeed, an IP may be resolved through a collaboration process involving internal watchers of an organization and, if necessary, with external watchers. This approach may also involve cooperation with other information workers. The question we are thus faced with is: how can we facilitate and manage collaboration in information seeking and retrieval to help solve an IP, with minimum cost and delay, while protecting the information patrimony of the company?

4.3. From information retrieval to CIR

4.3.1. Information retrieval

Information retrieval, as defined by Fidel et al. [FID 01], may be interpreted in a broader sense to include processes such as problem identification, analysis of information needs, query formulation, retrieval interactions, evaluation, presentation of results, and applying results to solve an IP.

An issue that is ever-present in the domain of information retrieval is that of relevance of the information found in relation to the information need of the user. This problem of relevance calls for reflection in the design and development of information seeking methods and tools. Possible reasons for the problem are as follows:

– incoherencies between an IP and its representation as a query;

– differences between terms used for indexing preexisting documents in information retrieval systems (IRSs) and the terms used by the user to express an information need;

– inability to capture the semantics of a user’s query;

– differences between the context of production of information and the context of final use [MAG 08];

– differences between the user’s mental model of an IRS and the functional model of the IRS;

– differences between user information behavior and user representation in IRSs.

The need to personalize the response of the IRS for each user, taking into account the variations in user preferences, is one of the developments that have emerged in the information retrieval domain. This gave birth to the idea of modeling users to better satisfy their information needs [BUE 01a, DAN 03].

Another development in IR is collaborative filtering and recommendations. This idea is not only concerned with personalizing responses but also with using what others have already found and evaluated. This hypothesis supposes that what is judged to be relevant by some users might be relevant for another user working on a similar IP [BUE 01b, GOH 05, JIN 06].

In spite of these developments, the problem associated with the semantic interpretation of user information needs expressed as a search query is still in want of a solution. For example, Google (www.google.fr) has introduced a system of automatic suggestion of search terms to users while they are typing their query. This constitutes progress on the part of Google, but potential problems arise from the inadequacy of a keyword or free text query to effectively represent user information need. If the typed query is a false representation of the user’s information need, the suggestions provided by the search engine might be considered to be an edifice built on a faulty foundation. A similar example of collaborative recommendation is proposed on the amazon.com (www.amazon.com) Web site, where a user searching for a book receives, as a response to his/her query, a list of suggestions based on books that were bought by other users who had formulated a similar query in the past. These developments are based on an approach of inferring users’ information needs based on their activities, profile, and queries through the use of algorithms.

From the cognitive viewpoint of information retrieval, we believe that it is, in fact, impossible to dissociate the problem of satisfying user information needs, and that of the relevance of information found in relation to the IP from the user’s understanding of the problem, and his/her level of knowledge of available information retrieval methods (or systems). This problem is strongly linked to the cognitive capacity of the user. This cognitive lack may be made up for by collaborating with other users, implying the notion of explicit collaboration in information retrieval and seeking [FOL 09, ODU 09, SHA 09].

4.3.2. Collaborative information behavior

CIR is a research domain that interests various disciplines including information and communication sciences (ICS), computing, linguistics, cognitive psychology, and artificial intelligence. Each discipline tackles the subject using its own set of rules. In ICS, the starting point is to study the information behavior of users to understand their information practices, before designing and developing IRSs and tools.

We will extend Wilson’s nested model of information behavior [WIL 99] to explain the three associated concepts involved in CIS. These concepts, as shown in Figure 4.1, are collaborative information behavior (CIB), CIS, and CIR.

In the outer circle of Figure 4.1, we find CIB, which concerns all information practices or behaviors of users seeking information in a collaborative manner, generally in more or less digital environments. CIB therefore includes CIS and CIR. CIB includes both active and passive information seeking and also includes the avoidance of information.

The second level, CIS, designates the context in which a group of users is found at the moment of choosing relevant information sources to respond to their information need (in this case, a shared information need).

The core, CIR, concerns the context of querying an IRS, that is, the collaborative formulation of queries and the collaborative evaluation of the relevance of the information found.

In the context of CI and watch activities in particular, observation of user behavior in collaborative information search situations shows that:

– users always find themselves in a loop made up of their IPs, information sources, and other users;

– users make use of several information systems in solving their IPs;

– users give more credit to information provided by an expert than to that found on an information system;

– users follow the evolution of an expert in their domain;

– users exploit the networks of other users they consider to be more experienced in their domains;

– users monitor the evolution of an information source, which they consider to be a good source of relevant information;

– users depend greatly on their social and professional networks in solving their IPs.

Figure 4.2 is an abstraction of these CIBs. Note that the trust factor is essential for determining with whom a user will collaborate or to what extent a user will express the totality of his/her IP and objectives.

Ihadjadene and Chaudirron conclude, in [IHA 09], that “information retrieval is at the heart of a process integrating search and navigation, collaboration, serendipity,² tagging, reading, or writing, all of which serve to complicate the modeling of information practices”.

4.3.3. CIS and retrieval

In the previous sections, we looked at information retrieval and CIBs, but we have not yet defined the word “collaboration” or the concept of CIS.

We define “collaboration” as the act of working jointly with shared objectives. Collaboration can be distinguished from cooperation, as in [LON 03]. In a context of “collaboration”, actors have a common goal, whereas in the context of “cooperation” this is not necessarily the case.

In Figure 4.1, we see that CIR is embedded in CIS, which means that CIR activities are subsets of CIS activities. We will henceforth stick to the term CIS to explain all information seeking and retrieval activities carried out in a collaborative manner. We thus define CIS as consisting of methods and systems for managing the collective activities of users in an information seeking and retrieval process to facilitate direct collaboration among them (the users) thereby enabling knowledge sharing [ODU 08a]. This definition highlights two important concerns in CIS:

– management of collective activities;

– facilitation of direct collaboration for sharing knowledge.

The aim of CIS is to share knowledge and to create new knowledge. In terms of knowledge sharing, the questions to ask are as follows: What knowledge should be shared in CIS? In what activities is this knowledge shared, and how may this knowledge be shared? In response to these questions, we will list the various forms of knowledge that may be shared [ODU 08b]:

– domain knowledge,

– competence in search methodology,

– system knowledge,

– knowledge of information sources,

– knowledge of collaborators.

4.3.3.1. Domain knowledge

Domain knowledge reflects the degree to which a user understands a search topic. It covers knowledge of facts, concepts, and terminology in a specific domain. High domain knowledge gives a user better access to relevant search results and gives the user the ability, through a richer set of concepts and terms, to formulate queries more effectively [SUT 98].

4.3.3.2. Competence in search methodology

The capacity of a user to plan his/her search depends on his/her competence in search methodology, which consists of search strategies, system selection, the use of operators, parentheses, truncation marks, formulation, modification, and expansion of search queries [ZHA 05].

4.3.3.3. System knowledge

Our knowledge of the functionalities of an IRS is very important in attaining our search goals. These functionalities may include a thesaurus, a list of keywords, a library of reusable queries, support for Boolean queries, visualizations of result summaries, visualization of the search process, and so on.

4.3.3.4. Knowledge of information sources

This knowledge relates to the ability of a user to find online databases, web resources, or IRSs that may be useful in solving a given IP [ZHA 05].

4.3.3.5. Knowledge of collaborators

All the types of knowledge mentioned above are possessed by users in differing proportions. The first question to ask when attempting to collaborate is with whom can we collaborate, that is, who are our prospective collaborators. Determining a prospective collaborator is a function of how best a user can mine the histories of other users and their models so as to determine their level of domain knowledge and other forms of knowledge mentioned above. The ability to mine and discover potential collaborators constitutes another form of knowledge that can be shared. We agree with the hypothesis that the best way of obtaining good information is to identify the person who possesses it.

To respond to the second question, “in which activities is knowledge shared?”, we will give an overview of these activities while at the same time identifying the knowledge to be shared in each of them:

– IP identification and clarification: the first activity in CIS is to identify the IP. Identifying a problem necessitates the clarification of the problem. Considering the fact that a user’s knowledge of his/her problem increases over time and that the problem may not be well understood initially, collaboration with another user may, therefore, help in clarifying the problem. The knowledge shared in this collaborative activity is domain knowledge.

– Articulation of the information need: the identification and clarification of an IP lead to the articulation of information need by formulating a search objective. Sharing domain knowledge is also very important in this activity.

– Choice of an information system to use: the articulation of search objectives leads to the choice of information sources to use for search. This activity requires sharing knowledge of information sources and competences in search methodology.

– Query formulation, reformulation, and clarification entail representing the information needs in terms of queries that are sent to the IRS. These activities require sharing domain knowledge, competences in search methodology, and system knowledge.

– Evaluation of results: this reflects the user’s judgment of the relevance of information found for their information need. This judgment is intrinsically linked to the user’s domain knowledge. Observation has shown that a document judged relevant today by a user may be judged irrelevant tomorrow by the same user with respect to the same IP and vice versa. This can be explained by the fact that a user’s knowledge of his/her problem increases over time. It also shows the importance of collaboration in evaluating the results of information search activities. For example, we might encounter a situation where a domain expert assists a user in determining whether a document is relevant or not to his/her information needs.

– Communication: communication among collaborators is required to share the knowledge in all the activities outlined above. This provides a response to the question of how knowledge can be shared. We will go into more detail with regard to this aspect in the following sections.

4.4. Facilitation and management of CIS

We will propose a conceptual framework and a communication model for managing CIS. The conceptual framework assists in understanding CIS and shows the four main aspects involved in managing CIS. The communication model is a model of the collaborative context for knowledge sharing.

4.4.1. The conceptual framework

Managing CIS involves the following four main aspects:

– communication,

– modes of collaboration,

– coordination of user interactions,

– management of knowledge involved in collaboration.

We are, effectively, considering a case where we communicate to collaborate, requiring coordination of interactions to successfully manage the knowledge involved in the collaboration.

These four elements are presented below.

4.4.1.1. Communication

For collaboration to be possible in the process of solving an IP, the actors involved must communicate. We will consider this aspect using the following questions:

– Why?: The “why” of the communication is the reason for the collaboration or the need for sharing, which can be seen as the objective of communication.

– What?: The “what” of communication concerns the communication object, whether audio, text, or video. These objects are knowledge expressed as queries, search results, annotations, and so on.

– How?: This aspect concerns the style of interaction that may be in either synchronous or asynchronous mode.

– Who?: The sender and receiver of the communications object (collaborators).

– When?: The date and time of communication, useful in contextualizing the expressed knowledge and in analyzing the evolution of users.

Communication is also considered a form of coordination of user interactions in the case of an information transfer − for example, the transmission of a document from one user to another.

4.4.1.2. Modes of collaboration

Two modes of collaboration may be found between actors involved in CIS:

– observation mode,

– interaction mode.

In observation mode, one or more users observe another user (expert or otherwise) carrying out search activities or attempting to solve an IP. Observation mode is represented by the graph in Figure 4.3a.

In interaction mode, two or more users attempt to resolve an IP conjointly. They share and exchange information and competences, each contributing to the resolution of the IP.

The graph in Figure 4.3b shows this mode of collaboration, with each user represented by a node

4.4.1.3. Coordination of user interactions

Coordination, in the context of CIS, is the management of dependencies between the information seeking activities of collaborating users. We support the notion expressed by Malone [MAL 94], who categorized possible dependencies between users into three general groups:

– flow dependency,

– sharing dependency,

– fit dependency.

Flow dependency occurs when the activities of a user become a resource for the activities of another user. For example, the query formulated by one user may become a starting point for another user with a similar information need.

Sharing dependency occurs when a resource is shared between two or more users.

Fit dependency occurs when the activities of two or more users must adjust to produce a resource. When two watchers work together on an IP, for example, their activities must be adjusted to fit together to produce a single resource.

4.4.1.4. Managing knowledge in collaboration

As mentioned in section 4.3.3, CIS involves knowledge sharing among users. This implies the need to manage the various types and forms of knowledge involved in CIS.

Knowledge to be shared includes queries, search results, annotations, dialogs between users, and shared documents. To manage this knowledge, we should consider the following:

– user modeling,

– knowledge acquisition,

– knowledge exploitation.

User modeling is carried out based on a user’s profile and activities. Knowledge acquisition entails capturing, encoding, and storing the different forms of knowledge expressed during collaboration, as set out above.

To exploit this knowledge, we will adopt the EQuA²te model [DAV 02]:

– Explore: to discover objects of the domain of study.

– Query: to access objects in the domain of study, from knowledge already acquired concerning the desired objects.

– Analyze: to obtain information with added value to discover the phenomena of the domain of study.

– Annotate: to create new knowledge. An annotation is seen as value added to information.

In the context of CIS, the EQuA²te model can be extended to include a syndication phase. After exploiting the knowledge base, a user may discover other users in his/her domain who constitute potential collaborators. The user may form a network with such potential collaborators, whence the term syndication.

4.4.2. Communication model for CIS

Successful collaboration depends on both collaboration technologies and a culture of openness among actors. Collaboration can, therefore, be said to be an interaction between technology and culture. Given these two factors, there is a need to model the collaborative context for interaction. Thus, for successful collaboration, integrating both collaboration technology and actors’ openness, there is a need for a knowledge sharing model, and a coordination mechanism for managing possible interactions in the collaboration.

From our previous studies of existing CIR and CIS systems and our own conception of CIS, we see CIS as communicating to share knowledge and acquire new knowledge [ODU 08a, ODU 08b, ODU 09]. The communication can take place in the interactions between two or more users passing by the IRS. It can also be manifested in the interaction between a user and the IRS as he/she formulates a query and sends it to the IRS, and the IRS provide him/her with results of relevant documents that correspond to his/her query. Summarily, we see every interaction process in CIS as a communication process because at any point in time, there is an information object being communicated from a sender to a receiver. The communication may be bidirectional as the receiver may also send an information object to the sender. We must also note that for every exchange, there is a context that necessitates the exchange as stated above (collaborative context). Based on this, we may model the process of communication in CIS. This model is known as the Communication model for Collaborative Information Retrieval (COCIR). The model is made up of four elements as shown in Figure 4.4.

As we stated earlier, we consider communication to be a means of collaboration and coordination; therefore, this model can be used to share knowledge and manage interactions among collaborators.

Each element of this model is made up of subelements which describe it. Both sender and receiver are users so they will have the same subelements. A user can therefore play both roles in a CIS session. Thus, when we make reference to the user, it may be either the sender or the receiver. The three elements of the COCIR model are made up of subelements presented in Figure 4.5.

4.4.2.1. User

Modeling users allows us to manage expressed knowledge during a collaborative session. It also allows us to link users with similar models. In modeling users, we combine the implicit and the explicit approach. Some information is explicitly provided by the user while other information is derived from his/her activities. Some of this information is static because it does not change, while other information is dynamic and may change with time. We combine the static and dynamic elements in representing a user. The user model contains information such as identity (surname and first name), date of birth, photo, and domain. The “location” subelement concerns the geographical location of the user during collaboration. This can be obtained from the internet protocol address of the computer used by the user. From the activities of the user (queries, annotations, consulted documents, etc.), we can obtain information on his/her domain knowledge, system knowledge, competence in search methodology, information source knowledge, and knowledge of collaborators. The “network” will contain information about users with whom the user had already collaborated.

4.4.2.2. Object

The object of communication represents all information exchanged among collaborators whether such information is expressed by them or comes from other existing artifacts. It is made up of four subelements: content, structure, format, and diffusion type. The content is the message to communicate. The message can be in different formats (text, audio, or video). The type of diffusion can be unidirectional or conversational. It is unidirectional when an information object is sent by one user to one or more others without expecting a reply. It is conversational when an object is sent to one or more users and produces a response. For example, when a user’s activity is automatically captured and sent to his/her partner during a direct synchronous collaboration for awareness purposes [ODU 10 p. 211–214], the diffusion type is unidirectional. Meanwhile, when two or more users engage in interpersonal communication during the course of the collaboration, the diffusion type becomes conversational.

Since one of the main aspects of CIS is the management of expressed knowledge, it is therefore necessary to capture and store this knowledge in the knowledge base. For knowledge to be stored, it must be structured. The subelement “structure” thus concerns the representation of information on the organization of the information object in the knowledge base. The structure varies according to the type of information object to be communicated.

4.4.2.3. Context

The third element of the COCIR model concerns the context of collaboration. Every exchange in collaboration applies to a specific context. In the framework of CI, the first subelement of “context” is the DP. As explained earlier, a DP can be translated into an IP. CIS activities are centered on IP. Thus, the second subelement is the IP, which can also comprise a set of IP that is, an IP may be broken down into various IPs. To model the DP, we adopt the Model for expressing a decision problem proposed by Bouaka [BOU 04, ODU 10]. Our major focus in CIS is the derived IP from the DP. Modeling the IP allows clarification of the problem to be solved and also allows collaborators to have a clear representation of the context initiating their interactions. IP will therefore contain subelements such as expression of IP, objective of search, keywords, information sources, domain, and temporal attribute [ODU 10].

This model can be used to facilitate CIS and the capitalization of knowledge expressed during the collaboration. An implementation of this model in developing a CIS system is seen in [ODU 10].

4.4.3. Application context

There are two possible cases of collaboration in information seeking. The first case corresponds to a situation where a user with an IP seeks another person with whom he/she can collaborate to solve the problem. The second case corresponds to a situation where a predefined group of users works together to solve a shared IP.

When discussing knowledge sharing in information seeking, three categories of knowledge should be captured to help users to:

– know other users better (potential collaborators);

– develop the necessary trust to begin collaboration;

– improve their social cognition;

– find already expressed knowledge that may be used to solve their problems;

– describe their metacognitive capacities.

These three categories of knowledge consist of:

– initial knowledge and competences,

– mental knowledge and competences,

– applied knowledge and competences.

Initial knowledge and competences are knowledge expressed and captured during the signing up of the user to the CIS system. They are deduced from the responses given by the user to a series of questions posed during registration.

Mental knowledge and competences are expressed during the definition, analysis, and clarification of the IP. They demonstrate the level of understanding a user has of his/her IP. They also describe how a user proposes to approach and solve his/her problem. In cases of collaboration where the problem is shared between several users, each contribution to the definition, analysis, and clarification of the problem via annotations and interpersonal communications falls into this category.

Applied knowledge and competences are derived from the information seeking activities and from interactions between users during resolution of an IP. They are expressed in the form of sources of information used, searches carried out, links visited, annotations made, users contacted, networks exploited, documents found, evaluations of these documents, and so on. This knowledge shows how users really succeed in solving a shared IP.

Let us begin by examining the first situation. A user connects to a CIS environment and defines an IP according to the attributes given in the problem definition interface. This definition relates to the user’s understanding of the IP. The user may progress to the level of information retrieval by connecting to an IRS and beginning to launch queries. However, since we are concerned with collaboration and faced with a situation where the user does not know with whom to collaborate, the user may attempt to exploit the collaborative knowledge base to find other problems similar to his/her own and the users who expressed such problems and who succeeded in solving them. This approach aims first at reusing potentially helpful preexpressed knowledge. However, the user may also want to identify users who possess, or have expressed, this knowledge. The goal is to enable direct collaboration to share tacit knowledge. This interest may be justified as follows, using an extract from a course given by Le Dantec [LED 07]: “The implementation of cognitive functions and processes is not the same for all subjects. However, at performance level, we regularly observe relative intra-individual, intra-task stability (particularly when tasks are of a similar nature), whereas the implementation of cognitive processes and functions does not necessarily work in the same way for the same subject in different tasks.”

Once the user has identified a potential collaborator, the user with the IP sends a collaboration request to the other user. If this user accepts the request, the IP becomes a shared problem for the two users and they begin a process of integrating and differentiating their understanding of the problem, aiming to obtain a shared understanding of the IP, and information sources to use.

From the instant when the problem is shared, the collaborative process that follows is the same in both cases discussed. The potential difference between the two is linked to the social cognitions of collaborators. We suppose that these cognitions are more developed in a predefined group where the members know each other.

Starting from a shared understanding of the IP, users begin to connect to information sources. This may be done in one of three ways:

– Each user connects to different sources then shares gathered information. User activities (navigations and searches) are also captured, providing the possibility for sharing them too.

– A user can observe another user in real time during navigation or searching. The user follows the activities of the other user and engages in interpersonal communication with this user to obtain explanations or clarifications about his/her actions when retrieving information. The first user may also express opinions to contribute to the actions of the other user. This is a learning process in collaborative working. For example, a user with considerable experience in the use of a specialized database may take the lead while others observe.

– All users (collaborators) formulate queries collaboratively and simultaneously. This is a more pronounced form of interaction. Users connect to the same IRS, formulate, clarify, and reformulate queries together. They evaluate the relevance of information together. They combine their queries and suggest keyword synonyms to use to find more relevant information. They engage in interpersonal communication and, at the same time, create annotations to express their knowledge on objects (a query, a document, a URL, a source of information, etc.).

Another interesting thing is to go from instant collaboration to user syndication, a process that creates networks of collaborators.

4.5. Collective information seeking scenario

To illustrate our approach for the facilitation of CIS, we will take the example of a student preparing a final dissertation on the subject “The impact of social unrest in the Niger Delta region of Nigeria on the international crude oil market”.

The student connects to an IRS (e.g. Google) and begins to look for information on the subject. He/she finds a large amount of information with a lot of background noise; his/her information need remains unsatisfied, as he/she is not prepared to sift through all of this information. In this example, the student uses the search engine alone by placing a search term.

Following our approach, he/she first defines the context as follows:

This definition shows the student’s mental knowledge of the problem. By sharing this problem with another user, the two users begin a differentiation and integration process to establish a shared understanding of the problem. In this example, the student, initiator of the problem, decides to connect with a user who knows the domain better than he/she does, and who has more knowledge of information sources. For the rest of this illustration, we will refer to the student as SU and to the second user as CU.

First, CU provides details concerning the domain: he/she gives a list of the states which make up the Niger Delta region of Nigeria. He/she then explains the forms of social unrest occurring in the region. Next, he/she highlights various possible categories of impact. This leads to a clearer understanding of the domain.

Following these clarifications, interpersonal communications take place between the two users through the CIS system so that they may harmonize their understanding of the problem.

Based on a shared understanding of the problem, SU and CU may begin navigating through information sources, and formulating queries to express the shared information need.

CU: “Now, I’ll formulate queries and you can see what happens: use of quotation marks, the need to break up keywords into several queries, etc.”

SU observes CUs activities in real time through the ISS using WISIWYS (What I See Is What You See) technology. In the course of this process, SU learns better ways of formulating queries in line with the document representation model of the IRS in question. Seeing the results found by CU, SU may express an opinion on the relevance of the documents found. Based on the knowledge acquired in the observation mode, he/she may also begin to launch queries.

At the end of the process, SU succeeds in acquiring knowledge of the domain (clarification of the region of Nigeria), and competences in search methodology (how to formulate targeted queries in Google). All the knowledge expressed during the collaboration process is capitalized for future reuse.

4.6. Conclusion

We have demonstrated that the information seeking is indispensable in the CI process aimed toward resolving DP. Several studies have shown that information seeking, which is both a cognitive and a social process, requires collaboration among users. This chapter has focused on managing collaboration activities of users in solving an IP derived from a DP.

CIS involves the management of collective activities and sharing of knowledge among users. The knowledge to share includes domain knowledge, competence in search methodology, system knowledge, knowledge of information sources, and knowledge of collaborators.

To facilitate knowledge sharing and to manage the collective activities of users, we proposed a conceptual framework for CIS and a communication model.

4.7. Bibliography

[BOU 04] BOUAKA N., Développement d’un modèle pour l’explicitation d’un modèle décisionnel: un outil d’aide à la décision dans un contexte d’intelligence économique, Doctoral Thesis, University of Nancy 2, Nancy, 2004.

[BUE 01a] BUENO D., DAVID A., “METIORE: a personalized information retrieval system”, in BAUER M., GMYTRASIEWICZ P.J., VASSILEVA J. (eds), User Modeling 2001, Springer-Verlag, Berlin and Heidelberg, pp. 168–177, 2001.

[BUE 01b] BUENO D., CONEJO R., DAVID A., “METIOREW: an objective oriented content based and collaborative recommending system”, Proceedings of the Twelfth ACM Conference on Hypertext and Hypermedia, pp. 310–314, 2001, available online at http://citeseer.ist.psu.edu/482403.html.

[DAN 03] DANILOWICZ C., NGUYEN H.C., NGUYEN N.T., “Model of user profiles and personalization for web-based information retrieval systems”, in ABRAMOWICZ W. (ed.), Knowledge Based Information Retrieval and Filtering from Internet, Kluwer Academic Publisher, Norwell (MA), pp. 121–136, 2003.

[DAV 02] DAVID A., THIERRY O., “Application of “EQuA2te” architecture in economic intelligence”, Information and Communication Technologies Applied to Economic Intelligence — ICTEI’2002, Ibadan, Nigeria, pp. 32–38, 2002.

[DAV 09] DAVID A., “Relevant information in economic intelligence”, in PAPY F. (ed.), Information Science, ISTE Ltd, London and John Wiley & Sons, New York, 2009.

[FID 01] FIDEL R., BRUCE H., PEJTERSEN A.M., DUMAIS S., GRUDIN J., POLTROCK S., “Collaborative information retrieval (CIR)”, The New Review of Information Behavior Research: Studies of Information Seeking in Context, vol. 1, pp. 235–247, 2001.

[FOL 09] FOLEY C., SMEATON A.F., JONES G.J.F., “Combining relevance information in a synchronous collaborative information retrieval environment”, in CHEVALIER M., JULIEN C., SOULE-DUPUY C. (ed.), Collaborative and Social Information Retrieval and Access: Techniques for Improved User Modelling, IGI Global, Hershey, PA, pp. 140–164, 2009.

[GOH 05] GOH D.H., FU L., FOO S., “Collaborative querying using the query graph visualize”, Online Information Review, vol. 29, no. 3, pp. 266–282, 2005.

[IHA 09] IHADJADENE M., CHAUDIRON S., “Electronic information access devices: Crossed approaches and new boundaries”, in PAPY F. (ed.), Information Science, ISTE Ltd, London and John Wiley & Sons, New York, 2009.

[JIN 06] JIN R., SI L., ZHAI C., “A study of mixture models for collaborative filtering”, Information Retrieval, vol. 9, no. 3, pp. 357–382, 2006.

[KAR 98] KARAMUFTUOGLU M., “Collaborative information retrieval: toward a social informatics view of IR interaction”, Journal of the American Society of Information Science, vol. 49, no. 12, pp.1070–1080, 1998.

[KIS 07] KISLIN P., Modélisation du problème informationnel du veilleur dans la démarche d’intelligence économique, Doctoral Thesis, University of Nancy 2, Nancy, 2007.

[KNA 07] KNAUF A., Caractérisation des rôles du coordinateur-animateur: émergence d’un acteur nécessaire à la mise en pratique d’un dispositif régional d’intelligence économique, Doctoral Thesis, University of Nancy 2, Nancy, 2007.

[LED 07] Le DANTEC C., Psychologie différentielle PSP2DF4_S4 — Interaction cognition/conation, Class given at the university of Rouen 2006–2007, available online at www.univ-rouen.fr/servlet/com.univ.utils.LectureFichierJoint?CODE=1172696611874&LANGUE=0.

[LON 03] LONCHAMP J., Le Travail Coopératif et ses Technologies, Hermes, Paris, 2003.

[MAG 08] MAGHREBI H., DAVID A., “Open system for indexing and retrieving multimedia information”, Actes du 36e congrès annuel de l’Association canadienne des sciences de l’information (ACSI), University of British Columbia, Vancouver, Canada, 5–7 June 2008.

[MAL 94] MALONE T.W., CROWSTON K., “The interdisciplinary study of coordination”, ACM Computing Surveys, vol. 26, no. 1, pp. 87–120, 1994.

[MAR 94] MARTRE H., Intelligence économique et stratégie des entreprises, La Documentation Française, Paris, 1994.

[MAR 95] MARTINET B., MARTI Y.M., L’intelligence économique, Editions d’Organisation, Paris, pp. 27–28, 1995.

[ODU 08a] ODUMUYIWA V., DAVID A., “Collaborative information retrieval among economic intelligence actors”, The Fourth International Conference on Collaboration Technologies (Collab Tech 2008), Wakayama, Japan, pp. 21–26, 2008.

[ODU 08b] ODUMUYIWA V., DAVID A., “A communication model for knowledge sharing in collaborative information retrieval”, International Conference on Knowledge Management (ICKM 2008), Columbus, Ohio, 23–24 October 2008.

[ODU 09] ODUMUYIWA V., DAVID A., “A user centered approach to collaborative information retrieval”, Proceedings of the 2009 International Symposium on Collaborative Technologies and Systems, Baltimore, Maryland, 18–22 May 2009.

[ODU 10] ODUMUYIWA V., La gestion de la recherche collaborative d’information dans le contexte du processus d’intelligence économique, Doctoral Thesis, University of Nancy 2, Nancy, 2010.

[REV 98] REVELLI C., Intelligence stratégique sur Internet, Dunod, Paris, 1998.

[SHA 09] SHAH C., MARCHIONINI G., KELLY D., “Learning design principles for a collaborative information seeking system”, Proceedings of CHI 2009, Boston, MA, 4–9 April 2009.

[SUT 98] SUTCLIFFE A., ENNIS M., “Towards a cognitive theory of information retrieval”, Interacting with Computers, vol. 10, pp. 321–351, 1998.

[THI 02] THIERY O., DAVID A., “Modelling of the user in strategic information systems and economic intelligence”, Journal of the Association for the Development of Software (ADELI), vol. 47, p. 12, 2002.

[WIL 99] WILSON T.D., “Models in information behaviour research”, Journal of Documentation, vol. 55, no. 3, pp. 249–270, 1999, available online at http://informationr.net/tdw/publ/papers/1999JDoc.html.

[ZHA 05] ZHANG X., LI Y., “An exploratory study on knowledge sharing in information retrieval”, Proceedings of the 38th Hawaii International Conference on System Sciences, Hawaii, 2005.

---

¹ Chapter written by Victor ODUMUYIWA.

1 “It is argued that the fundamental intellectual problems of information retrieval are the production and consumption of knowledge. Knowledge production is fundamentally a collaborative labor, which is deeply embedded in the practices of a community of participants constituting a domain” [KAR 98].

2 Serendipity is the fact of discovering something by accident and sagacity while in pursuit of something else (definition given by Walpole H. (1754) — source: www.intelligence-creative.com/354_serendipite_definition.html). According to [MAR 95], it is the art of finding the right information by accident. In competitive intelligence, “it allows us to identify the ‘blind spots’ of a strategy or unfounded but widely accepted beliefs, which may help a competitor or a new entrant to create a breach in a market”.