dc.description.abstract | Since the end of 20th century, the rapid developments of Internet and information
technology have changed everyone’s daily life and the way we manage business. With
knowledge booming and economy evolving, the information and data circulating within a
business entity can build up substantially and rapidly. The data accumulated through time
can be structural, non-structural, homogeneous or heterogeneous. It is essential for a
business to thrive in a global competition that internal information can be retrieved
effectively. This means that the internal search engine can locate and provide accurate
information to fulfill business requirements in a timely manner.
This case study looks into a world leading personal computer company. Currently, this
company focuses on ICT products in the global market. As the company expands,
information gathered from various functions in branch companies all over the world surged.
Now the company is faced with the challenge to consolidate and utilize the information
collected from various sources. Currently, the search engine in the company does not
always find the information that users needed. Often times, the system would provide
results that are not relevant at all. The search engine can barely identify the relevancy of the
subjects searched.
Therefore, this study proposes a more effective way for data searching and display. The
proposed method is more intuitive, and can prioritize the search results according to
relevancy, and refine the query results such that relevant results can be more easily
identified, and the frequency of downloading physical files can be drastically reduced.
Thus, this study enhances the competitive advantage by establishing a platform where
information can be effectively communicated and shared for operational purposes.
This study identifies several major problems such as the lack of relevancy, insufficient
summary data, and improperly prioritized results. To model a solution, literature on Internet
search and enterprise search were reviewed. We have also factored users’ behaviors in to
our proposal. The solutions include:
1. Improve the query terminologies by leveraging domain expert knowledge through
the Navigators system.
2. A fast preview function which can generate meaningful summaries.
3. Gather feedbacks from users to improve how search engine prioritizes results.
iii
In this study, we built a prototype to prove the concept, and hope to extend it with
more functional features. We analyzed the feasibility of the proposal from four aspects:
marketing, technical, legal and economic. Analysis shows the proposed scheme to be
feasible from these aspects.
| en_US |