股骨头坏死酸胀麻木:如何有效的搜索科技文献

HOW TO SEARCH FOR SCIENTIFIC DOCUMENTS IN AN EFFECTIVE WAY

By Prof. Lê ?ình Duy - Oct 14th, 2008

Tag: advanced skill, scientific documents, tutorial

As a students, teaching assistants, professors, etc you all have to search for information, news, survey, research field to support for your work everyday. This topic was written to share some experiment for a more effective way in finding scientific documents.

1. Why do we need to search for scientific documents?

• To find papers about what we are interested in, such as researchfields, new technologies, surveys, review state of the art, what havebeen done, what will be done, etc...
• To find certain course slides and talks in order to enrich and update our base knowledge.
• To find some illustrations for our presentations.

In scientific researching, we do not want to reinvent the wheel.Thus, before starting a project, we should search thoroughly to knowwhat have been done.

Crucial skills:

• Determine reliable source of documents.
• Determine the main keywords for searching.
• Gather documents.
• Organize documents.

2. Determine reliable source of documents: to leave out trivia or unnecessary documents. Some sources:

a) Digital libraries: websites which store papers from conferences, workshop, journals, etc... in digital format files. Recommended websites:

• IEEE Xplore (http://ieeexplore.ieee.org): papers in computer science, electrical engineering and other fields in natural science.
• ACM Digital Library (http://portal.acm.org/dl.cfm/): papers in computer science.
• ScienceDirect (http://www.sciencedirect.com/): journal papers in many fields,
• Springer (http://www.springeronline.com/): many journal papers, conference proceedings.
• Elsevier (http://www.elsevier.com): journal papers published by Elsevier.

b) Key journals, key conference proceedings: prestige journalsand proceedings of interests. Starters can rely on recommended rankinglist, or look for reliable source from references in papers and famouslabs publication.
For more information regarding journals and proceedings, please refer to:
http://www.procul.org/blog/2005/04/1...%8dc-may-tinh/
http://www.procul.org/blog/2005/07/1...1ng-phat-minh/

Recommended sources:

• Computer Vision - Pattern Recognition: http://iris.usc.edu/Information/Iris-Conferences.html
• Software Engineering: http://people.engr.ncsu.edu/txie/seconferences.htm
• Database: http://wwwhome.cs.utwente.nl/~apers/rates.html
• Networking: http://www.cs.ucsb.edu/~almeroth/conf/stats

c) Key labs, key persons: reading documents published by leadinglabs and professors will save us much time because their websites oftenhave adequate information about their newest publications. Finding keypersons requires intensive reading.

3. Tips on searching for key conferences and journals: take note of:

• Ones that are referenced many times in key papers.
• Ones that are target publication of key persons and key labs.
• Ones that have high rank in relatively sufficient charts (Assourav,Zaiane, Guofei, and NUS). Note that these charts are unofficial and notupdated.

4. Advice on searching for key papers:

• Key papers are usually referenced many times. They should be listed indigital libraries and websites such as CiteSeer, the authors’ homepage,courses and seminars. Papers we cannot find anywhere may not be keypapers and we can skip it.
• Normally, information about key persons, papers and keywords will berelated in a spiral graph. From keywords, we can find papers; fromthese papers we can find more accurate keywords and persons, and frompersonal homepages, we can find papers. In the course of searching, wewill gradually approach what we need.
• Type the whole name of a famous paper in Google search box to searchfor related works. The results will be papers which refer to it,correlative courses or a reading list. These clues will help us findthe final result faster.
• When to stop searching for key papers? Normally, we can stop when wehave 50-70 papers. We will then spend time scanning all of them toselect the most relevant papers. For every specific problem, there maybe only 20-30 state-of-the-art papers.

5. Searching according to purposes:

Before we starts searching, we should have clear and specific purposesin mind to quickly find the information we want. What we need is aplace to search (WHERE) and search terms (KEYWORDS). Recommendedaddress:

• Wikipedia: great place to learn about state of the art or unknown terms and technologies.
• Google: a few tips on using restrictions:
• Provide as many keywords as possible including file types.
• Use domain search to reduce the number of sites we will search. Forinstance the keywords will restrictthe search to only find results in US universities.
• DBLP (Digital Bibliography & Library Project) and all the websites mentioned above.

6. Archive, organize and search for documents:

Organizing what we have found is also important. Papers in the samecategory should be placed in the same folder. For each paper, we shouldname it with the corresponding information in this suggested style 

Besides, installing a desktop search engine such as Google Desktop will help us find the archived documents quickly.

7. Summarization:

• Searching skills can make searching more fast and efficient.
• We need to have a clear and specific purpose before searching.
• Build our list of reliable sources and update it frequently.
• Organize documents for easy access and searching.