Method and apparatus for providing search results in response to an information search request
AssigneeKoninklijke Philips Electronics N.V.
InventorsSrinivas Gutta, Vasanth Philomin, Miroslav Trajkovic
Life of this patent
Estimating the term…
Abstract
A system and method is disclosed for taking the results of a search (i.e., electronic documents), using the TFIDF of terms contained in those search result documents to quantify how related the search result documents are to one another, and organizing the search results based on those quantified relations. Using such a system and method, the search result documents can be displayed in such a manner that the relationships between the search result documents is evident to the viewer.
Claims19
A method for processing results of a search, wherein the search results are electronic documents comprised of terms, the method comprising the steps of:
- a) determining a document frequency (DF) of some or all terms in a global set of electronic documents, wherein the document frequency is a frequency of a term over a global set of documents;
- b) determining a term frequency (TF) of some or all terms in each document in the search results, wherein the term frequency is a number of times a term appears in a particular electronic document;
- c) determining a TFIDF (Term Frequency Inverse Document Frequency) for some or all terms in each particular electronic document in the search results by dividing the TF of step (b) by the DF of step (a);
- d) converting each electronic document in the search results into vector format whereby a corresponding vector entity is created for each electronic document, wherein the determined TFIDF values of step (c) are vectors;
- e) calculating similarity measures between the corresponding vector entities of step (d);
- f) using the calculated similarity measures to cluster the electronic documents in the search results by clustering their corresponding vector entities, whereby at least one cluster is formed;
- and
- g) selecting a representative electronic document for each of the at least one cluster by selecting a centroid vector entity in each cluster.
A method for processing results of a search, wherein the search results are electronic documents comprised of terms, comprising the steps of:
- determining a TFIDF (Term Frequency Inverse Document Frequency) for each term in each electronic document in the search results, wherein the TFIDF is a term frequency divided by a document frequency, wherein the term frequency is a frequency of a particular term in a document and the document frequency is a frequency of the particular term over a global set of documents;
- determining a similarity measure between electronic documents in the search results, wherein said similarity measure is based on the determined TFIDF values;
- and
- organizing the search results according to the determined similarity measures.
A system for processing results of a search, wherein the search results are electronic documents comprised of terms, the system comprising:
- a means for determining a document frequency (DF) of some or all terms in a global set of electronic documents, wherein the document frequency is a frequency of a term over a global set of documents;
- a means for determining a term frequency (TF) of some or all terms in each document in the search results, wherein the term frequency is a number of times a term appears in a particular electronic document;
- a means for determining a TFIDF (Term Frequency Inverse Document Frequency) for some or all terms in each particular electronic document in the search results by dividing a TF of each particular term by a DF of the particular term;
- a means for converting each electronic document in the search results into vector format whereby a corresponding vector entity is created for each electronic document, wherein the determined TFIDF values are vectors;
- a means for calculating similarity measures between the corresponding vector entities;
- a means for using the calculated similarity measures to cluster the electronic documents in the search results by clustering their corresponding vector entities, whereby at least one cluster is formed;
- and
- a means for selecting a representative electronic document for each of the at least one cluster by selecting a centroid vector entity in each cluster.
Dependencies are inferred from the claim text.