US 20090265373 A1
A tagging application (TA) is disclosed that comprises a web interface program (WIP) and a tagging program (TP). The WIP presents the facet tree by displaying elements of the facet tree selected by the user along with a known activation method for the user to select a facet element and to assign a value to the selected facet or facet element. The user may proceed to manually tag each facet and facet element by selecting a value representing the facet element to be tagged. The user may also eliminate manual selection of many of the facets and facet elements by choosing a tagging scenario and then applying the tagging scenario. Once the user has selected a tagging scenario, and has indicated that it is to be applied to the facet element, the TP takes the selected value and applies the selected tagging scenario to populate the tagging for the user in accordance with the tagging scenario. Once the tagging has been populated, the tagging program returns the tagged facet tree to the user at the user interface. Explanatory notes may be provided to the user along with the tagged tree.
1. An apparatus comprising:
a computer having a first storage electronically connected to a second storage;
a tagging program in the first storage; and
a facet tree database, a rule database, and an authorized end-user tagging scenario database in the second storage;
wherein responsive to a first end-user selection of an element of a facet tree and a second end-user selection of an authorized end-user tagging scenario from the authorized end-user tagging scenario database, the tagging program populates causes a processor of the computer to populate a plurality of tags in the facet tree in accordance with a selected facet and in accordance with the authorized end-user tagging scenario, and to return a tagged facet tree to an end-user authorized to access only an authorized portion of the facet tree.
2. The apparatus of
3. The apparatus of
4. The apparatus of
5. The apparatus of
6. The apparatus of
7. The apparatus of
8. The apparatus of
17. A computer program product comprising:
a computer readable medium;
a web interface program stored in the computer readable medium; and
a tagging program stored in the computer readable medium;
wherein the web interface program and the tagging program contain a plurality of instructions adapted to cause a processor of a computer to perform steps comprising:
accessing a facet tree database, a rule database, and a scenario database; responsive to a user selection of an element of a facet tree from the facet tree database, and a scenario from the scenario database, applying tags to the facet tree in accordance with the scenario and the element; and
responsive to the scenario being an authorized user tagging scenario, allowing the user access to access only an authorized portion of the facet tree.
18. The computer program product of
starting with a leaf node and populating selections upward to a root;
starting a root node and populating downward to all child nodes;
populating upward along all elements of a given path; populating downward along all elements of a given path;
selecting all peer elements of a given node;
starting with a given, non-leaf node and populating upward along a path; and
starting with a given, non-leaf node and populating downward along a path.
The present invention is related to the subject matter of U.S. patent application Ser. No. ______, Attorney Docket number AUS9200400949US1, incorporated herein by reference.
The present invention is directed to database access in general, and to a method and apparatus for scenario based automation of tagging in a faceted tree in particular.
Vast amounts of digital information are available to users through the interconnection of computers and storage by the Internet. Indeed, traditional paper records can be scanned and retained electronically. Emails or instant messages between a company's engineers, planners, or financial officers can have the same or greater significance as formal memorandums or journal articles in a knowledge database. Although available in great quantity, digital information must be accessed in order to be used efficiently and profitably. Typical search methods use queries consisting of a Boolean combination of words and phrases which return documents containing words or phrases that match the search query. Imprecision and ambiguities in the words and phrases affect the effectiveness of searches. Refining a search takes time and skill and a user's search capability usually improves, with practice and experience.
However, businesses need fast and reliable search capabilities that are not dependent on a user's skill and experience. Speed and reliability are improved by a search capability that precludes a null result. A null result occurs when desired information is not present in the data base, or the desired information is present, but the search term fails to locate the information. Speed and reliability are improved by a search that is controlled so that users are logically directed to the sought after information without distractions by non-relevant information. Thus the record count of documents responsive to a search should never be zero, and should include only documents that are relevant to the search.
In addition to speed and reliability, businesses need a search capability that can restrict accesses to its total amount of digital information. For example, not all employees have a need to access all information. Trade Secret information, while part of the total universe of a business's digital information, must have access limited only to those with proper authorization. Moreover, one division of a corporation may not need access to another division's information. Employee files including health records and financial data must be protected in accordance with various State and Federal laws. Thus the need for information access with speed and reliability is affected by equally important needs to restrict access where appropriate.
Personnel responsible for organizing the universe of a businesses' digital information have found that a controlled search can be achieved by a faceted search using a facet tree. A facet tree is a categorization scheme for hierarchially structured data and content records where the facets and facet elements of the tree guide the search, and a posting list of records is associated with each facet element. A facet is a node in the hierarchial network and a facet may have any number of facet elements. The total number of records in a facet is the sum total of all records in the posting lists of the facet elements beneath the facet in the facet tree.
A faceted search is conducted by a faceted search engine. Selection of a facet (a word or phrase) causes the faceted search engine to present the next set of facets, and so on along the facet tree until a final facet is reached and the facet elements are displayed. When the user selects one of the facet elements, the documents on the associated posting list are displayed. The relationship of the facets to each other and of the facets to the facet elements is created by tagging. In general, a tag is one or more characters attached to a set of data that contain information about the set, including its identification. By tagging is meant assigning a value to a facet so that it is identified with another facet in a hierarchical relationship and assigning a value to a facet element so that it is identified with a particular facet.
Current approaches to faceted search are based upon the tagging of data and facet elements in a strict parent/child relationship. Using a faceted search engine, end users choose one facet or facet element at a time. Document or data record counts are displayed in conjunction with the facet elements, such that end users are aware of the total number of records and also how many data records exist at a given facet element within a facet. The total number of records means the number of records for all of the combined facets and facet elements that have been selected.
Faceted search has the advantage of eliminating null results when end users are searching for records across many combined dimensions. To do this, data records that are tagged to a facet element are also made available at facet levels. By way of example, a user might use a faceted search engine to locate content in geographical, industry, and technology facets by choosing “Europe,” “Retail,” and “Supply Chain” as facets. When one does this, all of the records in posting lists identified with the facets of Europe, Retail, and Supply Chain are aggregated and returned. The total number of records returned is the set intersection of all of the data records rolled up to Europe through Retail and Supply Chain.
Another form of faceted search is a conversational search. In a conversational search, end users are presented with a series of questions derived from or tagged to a facet hierarchy, and the answers given by the end users result in faceted responses rendered by a conversational search engine. The response may include search engine results at the record level finding, for example, documents, titles, abstracts, and other information about those documents. The response may also include navigation from the facet that the question was asked about, to a succeeding facet or facet element. In a conversational scheme, the semantic content of the questions matters, and so the order in which the questions are presented is important.
Current methods for tagging of content and data records for faceted and conversational search are tedious, manually intensive, and prone to error. When content is tagged by content providers, authors, or other editorial roles, it involves human selection of elements from a faceted categorization scheme. The manual tagging solutions currently offered by content management providers require selection of single elements in combination. This results in a much higher number of mouse clicks, and a greater amount of time for tagging each record or document to be presented via a faceted or conversational search rendering engine. Tagging documents manually is often cost-prohibitive because of the number of steps required to select each element in a number of deep hierarchical structures.
Businesses have a need to apply policies to tagging to meet requirements for restricted access for security and legal reasons, but also to apply policies to tagging to promote the efficiency of providing results to end users. Automated tagging algorithms exist such as the Eureka algorithm and software from IBM Research. Other automated tagging algorithms are known. For example, United States Patent Application 2002/0091661, entitled “Method and Apparatus for Automatic Construction of Faceted Terminological Feedback for Document Retrieval” (the '661 application), discloses a method for extracting key informational concepts or “facets” from a collection of documents. In the '661 application, facets are chosen heuristically based on ‘lexical dispersion. As used in the '661 application, lexical dispersion is “a measure of the number of different words with which a particular word co-occurs within such syntactic contexts.” The facets and their corresponding lexical constructs provide a structured summary of the contents of a result set.
However, these automated solutions rely on factors such as word frequency in a document and cannot tag documents in accordance with business rules. In order to tag data records and documents in accordance with a stated policy, a person must manually perform the tagging according to specific business rules. While algorithms are capable of approximating human tagging based on semantic categorization, they cannot perform tagging in accordance with policies. Additionally, each time a tagging policy changed, the algorithm would need to be changed. Therefore, although labor intensive, human tagging is necessary to make the policy distinctions that cannot be made by a machine.
What is needed is a solution that allows for rapid selection of tagging facets and facet elements in a facet tree by eliminating repetitive user actions so that a user can quickly tag data records, documents, or collections of documents to a given facet tree via navigation of the elements in a tree format for a given facet.
The invention that meets the needs identified above is a tagging application (TA) comprising a web interface program (WIP) and a tagging program (TP). The TA provides instructions to a server connected to storage containing a facet tree database, a scenario database and a rules database.
The WIP comprises a user interface and a set of servlets. The user interface comprises one or more Java Server Pages (JSP's). The servlets read the facet element to be displayed from the facet tree database, acquire any associated rules for tagging scenarios based upon the particular facet element from the rules database, and acquire any associated rules based on end user characteristics from the rules database. The servlets apply the rules to the facet tree before presentation to the user on the user interface so that end users who are to tag data and content records only see the portions of the facet tree they are authorized to see and only have available the tagging scenarios that are applicable to their situation. The servlets make the applicable tagging scenarios available to the user at the user interface.
The WIP presents the facet tree by displaying elements of the facet tree selected by the user along with a known activation method for the user to select a facet element and to assign a value to the selected facet or facet element. For example, the user may expand the facet tree by clicking on a facet or facet element or by using a toggle button. The user may then assign a value to a facet or facet element by clicking a checkbox, or using any other known method. The user may proceed to manually tag each facet and facet element by selecting a value representing the facet element to be tagged. The user may also eliminate manual selection of many of the facets and facet elements by choosing a tagging scenario and then applying the tagging scenario.
Once the user has selected a value for the facet element to which a scenario is to be applied, the user accesses the available tagging scenarios by activating a user interface control such as a selection button, a drop-down list, or a link. Alternatively, the user may navigate to a specific JSP for a given tagging scenario. Once the user identifies the available tagging scenarios associated with the selected element, the user selects a tagging scenario.
Once the user has selected a tagging scenario, and has indicated that it is to be applied to the facet element, the TP takes the selected value and applies the selected tagging scenario to populate the tagging for the user in accordance with the tagging scenario. Once the tagging has been populated, the tagging program returns the tagged facet tree to the user at the user interface. Explanatory notes may be provided to the user along with the tagged tree. The tagged tree can be saved, modified according to a different tagging scenario if one is available, canceled, or saved in draft format for review or other workflow. Moreover, the TA can be combined with an automated tagging program to allow users to validate or invalidate selections made by one or more algorithms or standardized tagging schemes.
These and other objects of the invention will be apparent to those skilled in the art from the following detailed description of a preferred embodiment of the invention.
The novel features believed characteristic of the invention are set forth in the appended claims. The invention itself, however, as well as a preferred mode of use, further objectives and advantages thereof, will be understood best by reference to the following detailed description of an illustrative embodiment when read in conjunction with the accompanying drawings, wherein:
The principles of the present invention are applicable to a variety of computer hardware and software configurations. The term “computer hardware” or “hardware,” as used herein, refers to any machine or apparatus that is capable of accepting, performing logic operations on, storing, or displaying data, and includes without limitation processors and memory; the term “computer software” or “software” refers to any set of instructions operable to cause computer hardware to perform an operation. A “computer,” as that term is used herein, includes without limitation any useful combination of hardware and software, and a “computer program” or “program” includes without limitation any software operable to cause computer hardware to accept, perform logic operations on, store, or display data. A computer program may, and often is, comprised of a plurality of smaller programming units, including without limitation subroutines, modules, functions, methods, and procedures. Thus, the functions of the present invention may be distributed among a plurality of computers and computer programs. The invention is described best, though, as a single computer program that configures and enables one or more general-purpose computers to implement the novel aspects of the invention. For illustrative purposes, the inventive computer program will be referred to as the “Tagging Application (TA).”
As used herein, the term “inverted index” shall mean an index of words or phrases where each word or phrase in the index is associated with a list of texts, and each word or phrase is also associated with the location within each of the texts where the word or phrase occurs, so that when the inverted index is accessed by a search engine, each word or phrase in the index can be located and the location of the word or phrase in each of the texts containing the word phrase can be displayed. As used herein, the term “tagging” means assigning a value to a word or data record to (1) select the word as a facet or a facet element, (2) to link the facet or facet element to another to facet or facet element, or (3) to link a data record to a facet element. As used herein, the term “posting list” means one or more values assigned to a data record and linked to a facet element where the list of values will be displayed when the facet element is activated, and each of the data records will be displayed when the value in the posting list is activated. As used herein, the term “facet rendering engine” means a program that will display a file or inverted index that has been tagged as a facet tree.
The TA is described below with reference to an exemplary network of hardware devices, as depicted in
WIP 230 and TP 220 typically are stored in a memory, represented schematically as memory 200 in
Tagging Application (TP) 220 interacts with a number of databases and files located on a representation of remote storage 120 depicted in
When the user has selected a tagging scenario and indicated a desire to apply the selected tagging scenario, TA 500 applies the scenario (518) and populates the facet tree with appropriate tags (520). The user reviews the tagged facet tree (522), and selects a disposition of the tagged facet tree (524). If the user wants to tag another facet tree (526), TA 500 goes to step 506, and if not, TA 500 stops (528).
A preferred form of the invention has been shown in the drawings and described above, but variations in the preferred form will be apparent to those skilled in the art. The preceding description is for illustration purposes only, and the invention should not be construed as limited to the specific form shown and described. The scope of the invention should be limited only by the language of the following claims.