Documentation

Right truncation

“Truncation“ means that abbreviated terms are used for searching in a database or in an information retrieval system. With simple searches in the central reference database, an automatic right truncation is carried out for each search term: program also finds programming OR programmer.

Case sensitiveness

The system is not case-sensitive: CoMpUtEr searches for computer.

Umlauts

Umlauts are automatically romanised: Müller searches for müller OR mueller.

Search for phrases

Double quotes embrace phrases which are literally searched for: "Traveling Salesman" does not find salesman who is traveling.

Stemming

In information retrieval, “stemming“ describes a linguistic method of determining the common stem of morphologic variants of a word. io-port.net uses a Porter stemming algorithm for English texts. Example:
control, controls, controlled, controller, controlling, controllation all have the same word stem:control

Back            Top of page

(chargeable value-add service)

Search for literature in the central io-port.net reference database

Concatenation with "or"

You may concatenate several search terms within one entry field with the operator OR.

Concatenation with "and" or "and - not"

You may only concatenate search terms with AND or AND NOT between different entry fields in the search mask. Select AND or NOT by clicking the triangle beside the Boolean Operator field.

Selection of search fields

For each entry field in the search mask, you may determine which field in the database you would like to link it with. Click the triangle beside the field name to select a field. You may also select the same field type several times, if you want to link the terms in these fields by AND or NOT.

Right truncation

Right truncation is available for each search term. Set the checkbox accordingly

Case sensitiveness

The system is not case-sensitive: CoMpUtEr searches forcomputer.

Umlauts

Umlauts are automatically romanised: Hügel searches for hügel OR huegel.

Search for phrases

If an input field contains more than one word, the entire content is interpreted as one phrase which is literally searched: Traveling Salesman does not find salesman who is traveling. If this is not desired, and the words are supposed to be concatenated by "and" instead, the words have to be distributed over various input fields.

Back            Top of page

List of hits sorted according to a predefined ranking. Title, authors and source fields are displayed (see detailed display)

Internet document

This document is only available electronically via an URL. Its quality could not be checked.

Save query for alert service

(chargeable value-add service)

You may save your search query and activate an alert service. You will then be notified at every database update if there are new results matching your predefined query.

Back            Top of page

(chargeable value-add service)

Title

Title of the publication

Authors

List of authors and publishers. An author’s publication list can be selected by clicking on the author’s name.

Source

Information on the source, e.g., (for journals:) journal title, ISSN; (for other publications:) publisher, ISBN; (for conferences:) conference title, venue and date

Internet document

This document is only available electronically via an URL. Its quality could not be checked.

Full-text

Access to the full-text of the selected document e.g. link to a full-text document stored on io-port.net. Optionally, a link to a full-text supply service or to the user's OpenUrl link resolver can also be displayed (see FAQs for technical details).

External links

Links cited in the document.

Export

Export of the document in BibTEX or SML format.

Similar titles

A special algorithm finds documents with similar titles stored in the database. Priority is given to words contained in the title which are not frequently mentioned in the database. Both title and document type are displayed.

Back            Top of page

(chargeable value-add service)

Only available for search queries in English language.

The Query Refinement Tool, developed within the scope of the SemIPort project, helps users create an optimum search query. (Only usable for search queries in English language!)

The Query Refinement Tool is a semantics based method to optimize a search query step by step by adding relevant terms to it. This is particularly useful for searches in large, unstructured data stocks.

First, the user formulates his search query. In order to refine it, he may then activate the query refinement tool, which will generate two new terms: “modifications” , i.e., terms related to the initial search term, and “specialisations” , i.e., terms which define the search term more closely (see below).

Value-add

Searching in a web portal is often time-consuming and does not always yield optimum results. The reasons are:

1. Users do not know exactly what they are looking for.
2. Users do not know which information is available.
3. Users only enter short queries (1 up to 3 words)
4. Users only read top-ranked results

Therefore an efficient knowledge retrieval system is needed, providing semantic technologies to support the formulation of the search query. The Query Refinement Tool perfectly suits this purpose.

The benefits to the user are:

1. quicker, more efficient search
2. better exploration of the search area
3. better, more relevant results
4. optimized search query, not based on statistics but supported by semantic technologies

Background information

The Query Refinement Tool is based on the three-step Librarian Agent Query Refinement Process:

1. ambiguities in the query are identified
2. recommendations to optimize the query are given
3. recommendations are ranked and clustered

The information theory knows two general relations to explain the meaning of a term: “modification” and “specialisation”. Therefore, the Query Refinement Tool generates two types of recommendations:

1. "Modification" suggests related terms to be added to the search term, e.g., "risk management" and "risk assessment" to "risk"
2. "Specialization" suggests to more closely define the search term, e.g., to use "operational risk" or "credit risk" instead of just "risk"

The refinements obtained by this method are not to be seen as just additional search terms but as a semantic broadening of the query.

Further reading:

On the Conceptualization of the Query Refinement Task
N. Stojanovic
Library Management, Emerald Group Publishing, Volume 26, Number 4/5, pp. 281-194, 2005

An Approach for Step-By-Step Query Refinement in the Ontology-based Information Retrieval
N. Stojanovic, R. Studer, Lj. Stojanovic
The 2004 IEEE/WIC Conference on Web Intelligence (WI 2004), Beijing, China, IEEE press

Responsible for the Query Refinement Tool

Dipl.-El.-Ing. Nenad Stojanovic
Institut für Angewandte Informatik und Formale Beschreibungsverfahren - AIFB
Universität Karlsruhe
76128 Karlsruhe
Germany

Back            Top of page

(chargeable value-add service)

The SemIPort component "Recommender" sorts results according to a previously defined working context that has been transmitted to io-port. The most relevant search results (documents) are mentioned first. The working context is generated and administered with the "Document Manager" tool.

Responsible for SemIPort Recommender

Dipl.-Inform. Eric Schwarzkopf
Deutsches Forschungszentrum für Künstliche Intelligenz
Stuhlsatzenhausweg 3
D-66123 Saarbrücken

Back            Top of page

io-port.net offers its users the possibility of creating and publishing their own publication lists (private or faculty-related) in the portal.

Publication lists can be published and thus made available to the public; non-published lists can only be viewed by their owner under "my publication list".

Published lists can be accessed by all users, but only registered users can create publication lists (chargeable service).

Publication lists can be created, uploaded and administered with the io-port.net Editor.

Back            Top of page