Patent Grant
7769741
The present invention is directed to information retrieval systems, and more particularly to an integrated searching/browsing retrieval system that combines the functionality of a full-text search engine with the flexibility of a browser.
As greater amounts of information continue to be stored on, and accessed by, computerized information services, mechanisms which enable a user to locate and review information of interest become a critical component of an information retrieval system. Typically, one of two different types of mechanisms is employed by users to access documents that are relevant to their needs or interests. One such mechanism is the hierarchical browser, which permits the user to view groups of documents and identify the documents within the context in which they are stored. If a document of interest is located, related documents can be easily found as well, because the user is provided with the ability to look in areas surrounding the document of interest to see other documents that are stored with it. Hierarchical browsers are essentially one dimensional, in that a single tree-like structure is provided for navigating among documents. This mechanism may be acceptable for single-user systems, in which the user authored the hierarchy. However, in multi-user systems, such as news databases, help lines, work group databases, and the like, not all users may find the single hierarchy to be cognitively intuitive, and therefore may have difficulty navigating through the hierarchy to find documents of interest.
The other type of retrieval mechanism, namely a full-text search engine, permits users to enter queries, and respond to those queries by returning document titles. These titles may be accompanied by relevance rankings or other indicators of the degree to which each document corresponds to the query. The document titles are typically presented to the user in the form of a list that may be sorted by relevance rankings, and do not provide any information regarding the context of the documents in the full database of information. Furthermore, they do not give the user the ability to browse through the spaces that are occupied by the listed documents, to locate related documents which were not selected as a result of the query. Such systems tend to work best for users who are looking for specific information, and they require a certain level of skill in constructing appropriate queries, since the results are highly dependent upon the terms which the user employs as an expression of interest.
It is an objective of the present invention to provide a document retrieval mechanism that offers the functionality of a full-text search engine, to enable a user to readily identify documents meeting that user's criteria, while at the same time offering the flexibility of a browser, to permit the user to locate related documents that may not match a specific query but which may nevertheless be of interest to the user.
In accordance with the foregoing objective, the present invention provides an integrated searching/browsing mechanism that enables users to construct information hierarchies that represent a cognitive framework for the organization of information, and to apply those hierarchies to any database of information. With this approach, the user is provided with the ability to navigate any available document database in a manner that is intuitive to that user. Furthermore, a user can combine two or more hierarchies to locate documents which match the criteria of both hierarchies, to further refine the collection of documents that are presented to the user. A significant feature of this approach is that the hierarchies can be constructed and applied independently of any particular document collection.
The information retrieval system of the invention is comprised of three main components. Dimension files represent information hierarchies constructed by users. A search and indexing system indexes information and returns lists of documents in accordance with the hierarchical structure defined in the dimension files. A user interface enables the user to specify the information to be retrieved, and provides the user with the ability to browse through various hierarchical levels of information. A system comprised of these components can be applied to a variety of different sources of information.
The system of the present invention enables users to view any collection of documents via a cognitive framework, as represented by browser hierarchies, which is useful, meaningful and comfortable to the users. It also permits the users to switch from one browser hierarchy to another, or to combine hierarchies in a single browsing experience. As a result, the user is able to browse over independent document collections, rather than being required to perform targeted searches that may limit the information presented to the user.
Further features of the invention, as well as the advantages achieved thereby, are described in detail hereinafter with reference to an embodiment illustrated in the accompanying drawings.
The following description of the invention is made with reference to specific embodiments, to facilitate an understanding of the features and principles of the invention. For instance, examples of search hierarchies are given with respect to categories of products in which a user might be interested. It will be appreciated, however, that these specific examples are merely illustrative of one application of the invention, and that the features of the invention have a broad range of applicability to a variety of different types of information, and different categorizations of that information.
An embodiment of an information retrieval system which is based upon the principles of the present invention is illustrated in the block diagram of
The client application 10 executes on the user's computer (not shown). Preferably, this application is stored together with the dimension files on a suitable computer-readable medium associated with that computer, e.g. a magnetic disk. The search and indexing system 14 could also be running on the user's computer, or on a different computer to which the user's computer is connected, such as a network server. The information sources 16 might also be stored on the user's computer, but more typically would be associated with one or more servers to which the user's computer is connected.
The dimension files 12 represent browsing hierarchies that define the relationship between various search criteria, in a manner that is comprehensible to the user. The hierarchies represented by the dimension files can be created by the users themselves, as described in detail hereinafter, or they can be preexisting hierarchies having a structure about which the user is knowledgeable.
The client application 10 provides the interface via which the user selects one or more browser hierarchies from the dimension files 12, as well as the information sources 16 that are to be examined for documents of interest. The search and indexing system 14 constructs an index of each of the information sources 16, in a known manner. The search and indexing system can be any suitable conventional type of search engine that is capable of accepting compound boolean requests. Once a user has selected one or more information sources to be searched and the appropriate browser hierarchies to be employed in searching the information, the search and indexing system 14 returns a list of the documents which meet the browsing criteria. These returned documents are then displayed to the user via the interface provided by the client 10.
Each dimension file 12 corresponds to a browser having a structure that represents various levels of information in a hierarchical relationship that is determined by the user's personal preferences regarding the organization and content of information. In a preferred embodiment of the invention, the user can construct a browser hierarchy by creating a series of empty, nested folders within the graphical user interface of a computer's operating system. One example of such a user interface is the Finder component of the Macintosh Operating System, distributed by Apple Computer, Inc. By means of such an interface, the user can create a series of folders that are nested within one another in a manner that corresponds to a desired hierarchy. For instance,
To construct this browser hierarchy, the user first creates an empty folder labeled “Market”. Within this folder, the user then creates three folders respectively labeled “Business”, “Education” and “Home”. The “Home” folder contains two further folders labeled “Family” and “Home Office”. Once the folder hierarchy has been completed, the highest level folder, in this case the “Market” folder, is imported into the information retrieval system, to create a corresponding dimension file. The importation of the folder hierarchy into the system might be carried out by means of a suitable command on the user interface, for example. In response to the selection of such a command, the information retrieval system examines the hierarchical structure of the folders, and creates a dimension file which defines this structure.
The dimension files contain terms which characterize content to be gathered within the hierarchies. For the particular example depicted in
The “Market” dimension file represents one approach that the user employs to categorize information. In a similar manner, other dimension files are created by the user to represent other approaches to information content that are intuitive to that user.
One of the significant advantages of the dimension files is that they are created independently of any particular source of information. This feature enables them to be portable, so that their inherent structure can be applied to information from a variety of different sources. As a result, the user can carry the client application 10 and the dimension files 12 to any suitable information source, e.g., a library, and use them to search documents in a manner that is comfortable to that user. Furthermore, because of their portability, the dimension file can be shared by several users, and distributed to all the members of a group or enterprise.
In the operation of the system, the terms at the various levels of the hierarchy are used to construct queries upon which the search system 14 operates to retrieve relevant documents from selected information sources. More particularly, the terms that the user employs in constructing the browser hierarchy, e.g., the names of the folders, form components of the query. The actual terms which are used in the query are dependent upon a level within the hierarchy that is selected by the user. If the user selects the lowest level in the hierarchy, only the term associated with that level is employed in the query. Conversely, if the user selects the highest level in the hierarchy, each of the terms in that selected level and all levels below it are employed in the search query. For instance, all of the terms can be joined with boolean “OR” operators, to retrieve all documents that contain any of the teems in the selected level and all dependent levels of the browser hierarchy. This operation is explained in greater detail hereinafter with reference to specific examples that are illustrated in
Once the user has selected the appropriate information sources, the hierarchical browsers to be used in the searching of those information sources are selected. The available browsers, which correspond to the dimension files, are listed in a window pane 22 on the right side of the interface. It can be seen that these browsers correspond to the dimension files illustrated in
The window pane 22 includes a suitable arrow button, or other equivalent form of control element, which permits the user to view the various levels of each browser hierarchy.
Once these selections have been made by the user, the documents within the chosen information source(s) which correspond to the browser selections are displayed in a search results window pane 24. These documents are selected by means of a query which is presented to the search and indexing system 14. Referring to the hierarchy illustrated in
The user can expand the search by selecting a higher level within the browser hierarchy. For example, if the user moves up one level to select the “Home” category, the search and indexing system 14 then institutes a search for all documents which contain, or relate to, the word “Home”, as well as the terms which appear in all lower levels of the hierarchy that depend from the “Home” level. In this particular case, the search and indexing system 14 would locate documents that relate to the terms “Home”, “Family” and “Home Office”. By means of this approach, it can be seen that the present invention provides an information retrieval system which enables the user to institute a targeted search, by selecting a particular level within the browser, but also provides the flexibility to locate related documents according to the hierarchical structure of the browser, simply by moving to different levels. The browser interface provides the user with a convenient mechanism to expand or contract the search query, while remaining within the context originally established by the browser hierarchy.
If the search results did not identify the particular types of documents the user was looking for, or if the user simply desires to take a different approach to searching the documents, it is possible to switch to a different browser from among those which are available. In this manner, the same source of information can be viewed from a variety of different perspectives.
To further refine the search results, the user can combine multiple browsers which organize the information in different manners. For instance,
In the example of
To summarize the general operation of the information retrieval system of the present invention, a brief functional overview is provided in the flow chart of
To further enhance the user's experience of browsing a document collection, as opposed to executing a targeted search, the information retrieval system can analyze a selected document to identify branches of the browsing hierarchy with which that document is most closely associated. This feature of the invention is explained with reference to the exemplary hierarchy illustrated in
Once a search is conducted in accordance with the selected level of the browser hierarchy and the search results are displayed, the user can select one of the documents in the displayed results. The contents of that given document are examined, to determine how many times each of the terms in the relevant hierarchy appear within the document. In effect, a reverse indexing approach is employed, in which the words in the document are compared against the terms in the browser hierarchy, to determine which of those terms appears most frequently. In the example of
The values that are associated with the levels in the hierarchy identify the concepts in the hierarchy that are most frequently mentioned in the selected document. These values can be displayed on the interface, to provide feedback to the user on the type of relevance associated with the document. From this information, the user can modify the selection within the browser hierarchy. For instance, in the example of
As an alternative, rather than display the results that are obtained through this analysis, the client application could automatically reposition the browsing hierarchy to the location that most closely corresponds to the selected document's location, as indicated by the computed values.
From the foregoing, it can be seen that the present invention provides an information retrieval system that allows users to employ any knowledge that they may have about the information they are looking for to construct a cognitive framework within which to locate that information. This framework, in the form of browser hierarchies, is independent of the actual information to be searched, and therefore can be readily distributed to various users and applied to different sources of information as desired. Furthermore, the user has the continual ability to switch among alternative browsers, combine multiple browsers, and create additional contextual hierarchies, to fit any particular situation. By using these hierarchies to locate potentially relevant documents, the user is able to browse in spaces around those documents that are organized in accordance with the cognitive framework established by the user.
It will be appreciated by those of ordinary skill in the art that the present invention can be embodied in various specific forms without departing from the spirit or essential characteristics thereof. The presently disclosed embodiments are considered in all respects to be illustrative and not restrictive. The scope of the invention is indicated by the appended claims, rather than the foregoing description, and all changes that come within the meaning and range of equivalence thereof are intended to be embraced.
This application is a continuation of application Ser. No. 10/406,439, filed Apr. 4, 2003, now U.S. Pat. No. 7,613,684, issued Nov. 3, 2009, which is a continuation application of application Ser. No. 10/150,086, filed May 20, 2002, now U.S. Pat. No. 6,574,620, issued Jun. 3, 2003, which is a continuation application of application Ser. No. 09/176,116, filed Oct. 21, 1998, now U.S. Pat. No. 6,418,429, issued Jul. 9, 2002.
| Number | Name | Date | Kind |
|---|---|---|---|
| 5642518 | Kiyama et al. | Jun 1997 | A |
| 6070164 | Vagnozzi | May 2000 | A |
| 6094649 | Bowen et al. | Jul 2000 | A |
| 6212512 | Barney et al. | Apr 2001 | B1 |
| 6236987 | Horowitz et al. | May 2001 | B1 |
| 6414429 | Watkins et al. | Jul 2002 | B2 |
| 6418429 | Borovoy et al. | Jul 2002 | B1 |
| 6480857 | Chandler | Nov 2002 | B1 |
| 6567812 | Garrecht et al. | May 2003 | B1 |
| 20030074369 | Schuetze et al. | Apr 2003 | A1 |
| 20060053142 | Sebbane | Mar 2006 | A1 |
| 20070203892 | Adams et al. | Aug 2007 | A1 |
| 20080263009 | Buettner et al. | Oct 2008 | A1 |
| 20090006427 | Veeraraghavan et al. | Jan 2009 | A1 |
| Number | Date | Country | |
|---|---|---|---|
| 20100063964 A1 | Mar 2010 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10406439 | Apr 2003 | US |
| Child | 12611252 | US | |
| Parent | 10150086 | May 2002 | US |
| Child | 10406439 | US | |
| Parent | 09176116 | Oct 1998 | US |
| Child | 10150086 | US |
