System and method for searching and retrieving documents by their descriptions

• US 20060020583A1
• Filed: 07/23/2004
• Published: 01/26/2006
• Est. Priority Date: 07/23/2004
• Status: Abandoned Application

• Current Assignee: Teleportal.Ru
• Original Assignee: Teleportal.Ru
• Inventors: Putilov, Alexandr V., Ishchenko, Vasily, Baranov, Alexey V.
• Application Number:
  US10/897,536
• Publication Number:
  US 20060020583A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 16/358 Browsing; Visualisation therefor
• CPC Class Codes: G06F 16/93 Document management systems

Abstract

Described is a method and system for searching and retrieving documents by their descriptions stored in databases and information resources with different document creation standards. The described method allows to decrease of the volume of the information displayed to the user terminal at a user'"'"'s request and decrease of intellectual efforts necessary to analyze the information obtained and come to a decision. The practical result is being achieved owing to the fact that all the homogenous documents from different databases are sorted by separate folders, the rating of each document is determined within a folder, hereafter the coincidence number of the characteristics of certain documents is counted in the different folders and the final rating of each document is evaluated taking into account the overlapping number, and then the documents are sorted by this rating and sent to the user'"'"'s computer.

Claims

1. A method for searching and retrieving information, comprising the steps of:

receiving and processing by a retrieval system a user request for retrieval of documents from at least one database;

sorting by the retrieval system of the retrieved documents based on subjects thereof;

creating a folder for each subject of the retrieval documents, each of the folders containing the sorted documents with the same subject;

storing each of the retrieval documents in a corresponding one of the folders for the respective subject;

determining document specifying characteristics of each of the sorted documents;

within each folder, determining a characteristic rating of each characteristic of each document stored therein;

determining a number of characteristics of each of the stored documents in a selected one of the folders that coincide with characteristics of the documents stored in the other folders;

determining a preliminary document rating using the characteristic rating of the document specifying characteristics of each sorted document;

calculating a final document rating of each of the sorted documents using the determined number of coinciding characteristics and a weighting factor of the database;

sorting the documents in accordance with the final document ratings; and

sending documents, sorted by the final document ratings, to the user.

2. The method according to claim 1, further comprising the steps of:

calculating a database rating for each database; and

storing the database ratings in the retrieval system.

3. The method according to claim 2, wherein the final document rating of the sorted (i-th) document is calculated according to the following formula:

$R_i=\underset{\begin{array}{c}j=1\\ x_{\mathrm{ij}}\ne <br><br>0\end{array}}{\overset{n}{\sum <br><br>}}<br><br>a_j<br><br>\frac{1}{x_{\mathrm{ij}}}+l_i+c_i,i=\stackrel{\_}{1,m}$

where x_i,j is a preliminary document rating of the i-th document in the j-th database;

a_i is a database rating of the j-th database;

l_i is a quantity of not equal to zero document ratings of the i-th document in all databases; and

c_i is a coincidence number of the different characteristics of certain documents in different folders.

4. The method according to claim 3, wherein the database rating of the j-th database is between 0.1 and 1.0.

5. The method according to claim 1, wherein the characteristics of the documents include authors, organizations, news, events, types of scientific and technical literature and patent documentation identified in the documents.

6. The method according to claim 5, wherein articles in bulletins, monographs, collections of works, proceedings of conferences and other scientific meetings are treated as different kinds of scientific and technical literature.

7. The method according to claim 2, wherein the database rating is determined based on a benchmark test.

8. A system for searching and retrieving information, comprising:

a request transmitter receiving and processing a user request for retrieval of documents from an information database;

a standards database in communication with the request transmitter, to provide data to aid in the processing of the user request;

a document integrator collecting the retrieved documents and storing the retrieved documents in a unified repository database;

a document sorter sorting the retrieved documents based on the subjects thereof and storing the retrieved documents in folders corresponding to the subjects in a folders database;

a characteristics processor determining document specifying characteristics, storing the characteristics in a characteristics database, and determining a characteristic rating of each document characteristic;

a reconstruction processor determining a number of characteristics of each of the documents in a selected one of the folders that coincide with characteristics of the documents stored in the other folders;

a rating calculator calculating a final document rating of each of the retrieved documents using the determined number of coinciding characteristics and a weighing factor of the information database and sorting the documents in a results database according to the final document ratings.

9. The system according to claim 8,

wherein the information database includes a plurality of databases.

10. The system according to claim 9, wherein a database rating is calculated for each of the plurality of databases.

11. The system according to claim 8,

wherein the rating calculator determines a final document rating of a (i-th) document according to the following formula:

$R_i=\underset{\begin{array}{c}j=1\\ x_{\mathrm{ij}}\ne <br><br>0\end{array}}{\overset{n}{\sum <br><br>}}<br><br>a_j<br><br>\frac{1}{x_{\mathrm{ij}}}+l_i+c_i,i=\stackrel{\_}{1,m}$

where x_i,j is a preliminary document rating of the i-th document in the j-th database;

a_j is a database rating of the j-th database;

l_i is a quantity of not equal to zero document ratings of the i-th document in all databases; and

c_i is a coincidence number of the different characteristics of certain documents in different folders.