Expert Documentation Approach for Knowledge Management in Online Forums

This research proposes an expert documentation approach for knowledge management in online forums based on opinion ratings of members. The ExpRank algorithm is extended from PageRank to consider both positive and negative ratings. Empirical assessment shows that ExpRank outperforms PageRank.

Knowledge Management
1

Abstract
There is no doubt in the fact that online forums are widely considered within a number of
organizational KM (knowledge management) practices together with virtual societies in order
to effectively share knowledge and viewpoints. Taking the above discussion into
consideration, this particular research intends to propose an expert documentation approach
that revolves around the judgment ratings of the members, mentioned on the online forum.
Further, in specific, this particular paper extends PageRank and proposes the ExpRank
algorithm, which considers both constructive along with undesirable agreement relationships
amongst online forum members. By making use of two datasets (relating to distinct product
segments i.e. music and books) gathered from a renowned product-review online site (i.e.,
Epinions.com), the outcomes of empirical assessment signify that the projected ExpRank
algorithm outpaces its benchmark approach (i.e., PageRank). Finally, results of evaluation
throw light upon the integration of undesirable agreement relationships could greatly enhance
the efficiency of expert recognition.
2

Introduction
To start with, online forums have been widely adopted within several organizational KM
(knowledge management) procedures along with virtual societies for the purpose of sharing
knowledge as well as perspectives (Wu et. al., 2006). Recognizing professionals within
particular sphere is vital for enhancing knowledge sharing along with availability via online
forums. Present expert recognition approaches could be broadly grouped into two key
concepts i.e. link-based and content-based (Nonaka et. al., 2016). Even though, the link-based
concept has illustrated its preeminence over content-based concept, it involves few
restrictions at the time when applied for recognizing specialists within online forums (Wang
et. al., 2006). Likewise, this particular study attempts to put forward an expert documentation
approach, which greatly depends upon the judgment ratings of the members, present within
an online forum. Further, in specific, this particular paper extends PageRank and proposes the
ExpRank algorithm, which considers both constructive along with undesirable agreement
relationships amongst online forum members.
A Review of Relevant Literature
There is no doubt in the fact that effective Knowledge Management has become most
important task for every company (Sanchez, 2016). The companies today should hold
complete knowledge about KM tools, KM techniques, KM models, KM processes,
knowledge sharing and trust, applying knowledge for innovation, new knowledge creation
and KM governance (Alavi and Leidner, 2001). In this particular paper emphasis has been
laid upon KM processes, KM techniques, new knowledge creation and KM models (Wu et.
al., 2006). In order to compete effectively within emerging knowledge-directed economy,
companies worldwide have taken up a number of initiatives, which aim at managing their
highly valued still impulsive asset i.e. knowledge. KM (Knowledge management) signifies
towards the methodical approach for generating, retaining, arranging, reusing, sharing and
lastly, assimilating explicit and/or tacit knowledge for the purpose of supporting the learning
procedures in the companies, thus resulting in enhanced organizational adaptability as well as
outcome (Wright, 2015). Additionally, the KM initiatives could be grouped into four
different sections i.e. developing knowledge repositories, enhancing access to knowledge,
improving knowledge atmosphere and lastly, dealing with knowledge like some asset
(Davenport & Prusak, 1998). Moreover, amongst them, knowledge access enhancement lays
emphasis upon enabling knowledge sharing amongst people, particularly from well-informed
3

people to others, for the reason that well-informed people could frequently provide answer to
questions and lastly, carry out required tasks needing exceptional understanding, experiences
and abilities (Wang et. al., 2013). Nevertheless, locating people having expertise or
knowledge (i.e., experts) for some particular requirement is frequently a complex job
(Davenport et al., 1998). For enhancing the knowledge accessibility, proficient recognition
systems, which could automatically recognize experts for some specific area are important
for knowledge admittance enhancement efforts (Wang et al., 2013).
Moving ahead, online forums are being widely utilized within several organizational KM
procedures along with virtual societies for the purpose of effectively sharing knowledge
along with thoughts (Nonaka, 2011). An online forum member could effectively share his/her
ideas in form of posts within the forum. Within few online forums, members could
effectively comment or respond to the posts put across through other members or also, could
rate as positive or negative to posts of other members. Therefore, along with the posts made
through other members, the connections amongst individuals also offer significant data for
effective recognition tasks (this involves recognizing specialists within specific areas) in
online forums (Wu et. al., 2006). Taking a step ahead, in reaction to the restrictions of the
prevailing link-based expert recognition tools, this particular paper attempts to propose an
expert recognition approach grounded upon opinion ratings provided through members
within online forums. In specific, the paper extends the PageRank algorithm, which is a
graph-built ranking tool frequently adopted within prevailing link-grounded expert
recognition methods, for developing an ExpRank algorithm that taken into consideration both
the positive as well as undesirable opinion ratings normally seen within online forums.
A Review of the KMS
There is no doubt in the fact that the present link-based expert recognition methods could be
operative at the time when adopted for the examination of email communications for the
reason that email exchanges within a company normally are conscious conducts impacted
through organizational standards as well as social links (Wang et. al., 2013). Thus, a person
who holds knowledge about a particular subject is likely to receive and respond to several
emails relating to the main theme (Hayes and Walsham, 2013). Therefore, edges modeling
email communications are considered through the prevailing ranking means as being positive
to the proficiency scores of people engaged. Nevertheless, within an online forum,
connection among people might not be positive always. For instance, in case if a person posts
4

his/her thoughts related to a particular subject on online forum. Other people might not show
consent with his/her views and respond with disagreement or provide negative ratings to the
post. The present link-based expert recognition methods don’t model negative edges within
their ranking devices and therefore, might not be efficient for expert recognition within online
forums. Additionally, in reaction to such issues, this particular paper proposes an expert
recognition approach based upon the opinion ratings amongst the people in online forums.
Further, in doing this, the actuality that PageRank exceeds or matches the HITS (authority)
algorithm in accurateness for expert recognition tasks on email communications is leveraged
(Dom et al., 2003). Thus, the PageRank algorithm is extended in this paper through enabling
it to take into consideration both negative and positive outlook ratings within online forums.
In reaction to the restrictions of the prevailing link-based expert recognition tools, this
particular paper attempts to propose an expert recognition approach grounded upon opinion
ratings provided through members within online forums. In specific, the paper extends the
PageRank algorithm, which is a graph-built ranking tool frequently adopted within prevailing
link-grounded expert recognition methods, for developing an ExpRank algorithm that taken
into consideration both the positive as well as undesirable opinion ratings normally seen
within online forums. The core of the suggested approach is the ExpRank algorithm that
considers the reply model (i.e., who retorts to whose posts) along with opinion ratings being
its inputs and creates a proficiency score for every person taking part in online forum.
Moreover, the opinion rating is emotion that differs greatly in extent and could be negative or
positive, of a person toward some particular opinion (or post) posted through other person
(Wu et. al., 2006). For instance, provided a particular rating scale (for example: strongly
agree, agree, somewhat agree, and disagree) a person who reacts to the opinion of some other
person with ‘‘strongly agrees’’ develops an opinion rating (in terms of this specific view).
Based upon the information present on the online forum, the suggested approach might
include an extra preprocessing stage, i.e., semantic relationship annotation. For instance, few
online forums just permit individuals to post views and respond to opinions of other
participants, thus being deficient of explicit opinion ratings. Moreover, within these forums,
opinion ratings may be understood, rooted within the answered messages. During such
situations, the proposed approach needs the semantic relationship annotation stage for
semantically annotating likely sentiment from the reaction to some particular post (Wu et. al.,
2006). In contrary, within other online forums (such as Yahoo! Answers), contributors could
show their emotions explicitly through voting on opinion posts submitted through other
individuals. During these situations, opinion ratings are available willingly and semantic
5

relationship annotation isn’t required. Few semantic relationship annotation approaches have
been devised and a large number of online forums are taking up voting mechanisms or
opinion rating because of the propagation of Web 2.0.
Critical discussion
The results of evaluation also propose the usefulness of undesirable agreement relationships
for sound expert-documentation. Moreover, this particular paper involved various limitations,
which require high research focus. First of all, for the purpose of evaluating the efficiency of
the suggested ExpRank algorithm as well as the benchmark approach (i.e., PageRank), one
must have a list of ‘true experts’ for individual dataset. Nevertheless, the majority of online
forums, does not choose and broadcast an expert list for every domain, making information
gathering highly limited.
For this reason, this paper employed datasets from one online forum (i.e., an acknowledged
product-review online site) for purposes of empirical evaluation. For enhancing the outside
generalizability of the results gathered during the study, it’s vital as well as preferred to
examine the suggested ExpRank algorithm by making use of datasets gathered from extra
online forums of different or similar kinds (for instance, discussion forums in companies).
Secondly, the paper extends PageRank to devise the ExpRank algorithm. During the coming
times, it will be quite fascinating to take on other ranking means (such as HITS authority) for
developing an expert recognition algorithm along with empirically match its efficiency with
the suggested ExpRank algorithm. Thirdly, it is assumed that the obtainability of opinion
ratings on online forums and as a result, fails to consider the semantic relation annotation
within the suggested approach. Additional research could involve developing a sound means
for semantic relationship annotation to ensure that the suggested ExpRank algorithm could be
adopted for online forums wherein explicit viewpoint ratings are unavailable.
Lastly, few members within an online forum might make an effort of inflating their
proficiency scores unprofessionally through making ‘‘fake’’ members as well as hiding such
pseudonyms for producing constructive opinion ratings for the posts (Ku et al., 2012).
Additionally, the presence of such deceitful members is expected to dampen the involvement
of normal individuals in online forums. As a result, detection of spam members becomes a
vital concern for online forums. Further, as with the suggested ExpRank algorithm, a likely
resolution for detecting spam member is by exploiting undesirable agreements. Even though,
6

ExpRank is developed for the purpose of recognizing experts, further study could extend
ExpRank for developing sound ways for detecting spam members.
Conclusion
To conclude, it can be clearly stated from the above discussion that online forums have been
widely adopted within KM procedures along with virtual groups for the purpose of sharing
opinions and knowledge (Wu et. al., 2006). Recognizing specialists within specific areas is
highly important for the sound knowledge sharing amongst the members if the forum.
Likewise, this particular paper proposed an expert recognition approach based upon the
opinion ratings provided through online forums members. In particular, the paper extended
PageRank and proposed the ExpRank algorithm, which takes into account both constructive
as well as undesirable viewpoint ratings. Moreover, by making use of two datasets associated
with distinct product groups (i.e. Music and Books) gathered from Epinions.com, the
empirical assessment outcomes exhibit that the projected ExpRank algorithm outdoes
PageRank.
7

References:
Alavi, M. and Leidner, D. E. (2001). Review: Knowledge Management and Knowledge
Management Systems: Conceptual Foundations and Research Issues. MIS Quarterly, 25 (1):
107–136.
Davenport, T. H. and Prusak, L. (1998). Working knowledge: How organizations manage
what they know. Boston: Harvard Business School Press
Davenport, T. H., DeLong, D. W. and Beers, M. C. (1998). Successful knowledge
management projects. Sloan Management Review, 39(2), 43–57.
Dom, B., Eiron, I., Cozzi, A. and Zhang, Y. (2003). Graph-based ranking algorithms for e-
mail expertise analysis. In Proceedings of the 8th ACM SIGMOD workshop on research
issues in data mining and knowledge discovery (pp. 42–48). San Diego, CA.
Hayes, M. and Walsham, G. (2013). Knowledge sharing and ICTs: A relational perspective".
In Easterby-Smith, M.; Lyles, M.A. The Blackwell Handbook of Organizational Learning
and Knowledge Management. Malden, MA: Blackwell. pp. 54–77.
Ku, Y. C., Wei, C. P., & Hsiao, H. W. (2012). To whom should I listen? Finding reputable
reviewers in opinion-sharing communities. Decision Support Systems, 53(3), 534–542
Maier, R. (2007). Knowledge Management Systems: Information And Communication
Technologies for Knowledge Management (3rd edition). Berlin: Springer.
Nonaka, I. (2011). The knowledge creating company. Harvard Business Review, 69 (6): 96–
104.
Nonaka, I., Krogh, G. and Voelpel, S. (2016). Organizational knowledge creation theory:
Evolutionary paths and future advances. Organization Studies, 27 (8): 1179–1208.
Rao, M. (2005). Knowledge Management Tools and Techniques. Elsevier.
Sanchez, R. (2016). Strategic Learning and Knowledge Management. Wiley, Chichester
Wang, G. A., Jiao, J., Abrahams, A. S., Fan, W. and Zhang, Z. (2013). ExpertRank: A topic-
aware expert finding algorithm for online knowledge communities. Decision Support
Systems, 54(3), 1442–1451.
Wei, C., Cheng, T. H., & Pai, Y. C. (2006). Semantic enrichment in knowledge repositories:
Annotating semantic relationships between discussion documents. Journal of Database
Management, 17(1), 49–66
Wu, B., Goel, V. and Davison, B. D. (2006). Propagating trust and distrust to demote web
spam. In Proceedings of the 15th international world wide web conference. Edinburgh,
Scotland
8

Wright, K. (2015). Personal knowledge management: Supporting individual knowledge
worker performance. Knowledge Management Research and Practice, 3 (3): 156–165
9