Search Modes

Objects in the Knowledge Base are connected to each other using various relationships, so that they create a semantic network. It is a graph structure. Searching in a graph structure can be performed in two ways:

by taking into account the graph structure and indicating in the query how particular types of objects should be combined - this is a graph search
by ignoring the graph structure and treating individual network elements as related texts, the text of the publication's description is "extended" by the text of the person's description and we perform a search in such extended texts - we then have a text search.

Graph search

Graph search allows you to search for objects connected to other objects, for example when you want to search for publications connected to a specific person using the author relationship. Graph search takes place when:

the system provides suggestions indicating a specific object, depending from the searching attribute, e.g. a person in the author attribute, journal in the series attribute, institution in the consortium attribute, etc., to be connected to the searched objects (an example is shown in Fig. 1 below),
the user will select one of the suggestions (figures 1 and 2) and attach the object specified in the suggestion to the question.

Fig. 1 Suggesting objects for graph search in terms of the relationship between paper and the journal

Fig. 2 Finding a person for graph search (when persn plays a role ofan author)

Graph search is characterized by the fact that the search is exact and includes only those objects that are actually connected to the selected object. For example, when we search for all publications authored by a specific person and we select in the Author field this specific person from the suggestions provided by the systen, we will obtain those publications that are connected by the author relationship with the indicated person.

Text search

You can perform a text search by entering text into any field. When the system attempts to provide you with a suggestion list you can ignore the suggestions by using the ESC key, then the entered text will be used for text search. For example, if we provide the text rybinski in the simple search (without selecting a specific person from the suggestions list, the system will search for publications where the text rybinski appears anywhere in the record. An example is illustrated in the figure below.

The text search syntax allows for complex formulas. The basic structures in the Knowledge Base are:

operator	syntax	an example and explanation

operator	syntax	an example and explanation
* - masking end of term	term*	energ* such a term results in retrieving all documents, where a term occurs starting with energ, e.g. energy, energetics, etc
conjunction (space or AND)	term1 term2 also the form as below is available: term1 AND term2	text mining the system will retrieve all the documents where both terms occur
disjunction (OR)	term1 OR term2	bozon OR mion the system will retrieve all the documents where at least one of the terms occurs
negation (AND NOT)	term1 AND NOT term2	bozon AND NOT mion in this case the relevant documents are those where bozon occurs and mion is missing
adjacency of terms (closing in quotations “ “)	“term1 term2”	“text mining” those documents will be retrievedwhere the terms text and mining are adjacent and in the same order
neighborhood	“term1 term2” ~3	“mining text” ~3 those documents will be retrieved where the terms text and mining are at distance not bigger than 3 words, and in the same order, e.g. like in the sentences: mining of text mining of the text

REMARKS

Negation occurs only in conjunction with the AND operator (AND NOT) and cannot start queries - it can only be used in the form term1 AND NOT term2.
Logical functions must be written in uppercase letters.