Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1998-06-24
2001-12-04
Homere, Jean R. (Department: 2177)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C707S793000, C707S793000, C707S793000, C704S009000
Reexamination Certificate
active
06327589
ABSTRACT:
FIELD OF INVENTION
The field of the invention relates to searching a file containing for topics and related terms and more particularly, to performing a search on a file having a format unsupported by a search engine and associating user queries to topics contained in the file.
BACKGROUND OF THE INVENTION
Software products often include online help files to assist users in performing tasks within the software. Users have become accustomed to executing these online help routines to assist them in performing tasks with the software. In more complex software environments, multiple help files may be used to provide the user with the requested help. In many cases, a search engine is used to search the help file, or files, and retrieve the relevant topics the user is requesting.
Search engines typically operate by the user first entering a query containing the topic in which the user desires help. The search engine then retrieves the help information related to the topic. Some search engines merely provide a user with a menu of help options, requiring the user to select from the available help options, and then retrieve the related help information based on the selected options. However, other search engines have become much more sophisticated by allowing a user to enter a query in a natural language. A natural language is a language that is structured to match the pattern of speech. For example, “How do I merge table cells?” is a query entered using natural language. These advanced search engines parse the query for important terms located in the query and use these terms to search the help files for related help information. This type of search is typically performed by using an indexing mechanism relating specific terms to the help information provided in the help files.
Even though the size of help files has increased to keep up with the increasing complexity of software and the number of tasks performed using the software, help files may not provide all the required help. Generally, online help files are not all inclusive and may leave out specific tasks. This may occur because the overall software community may only utilize a program function rarely, thereby justifying a software developers decision to leave out information on the function from the help file for performance reasons. This will, however, burden an individual user or office that may utilize the function often and desire to search for help on how to use the function. In addition, help files may not include specialized tasks created especially for the business using a general software product. In this case, the software developer fails to include the specialized task within the help file because he or she is unaware of the task.
Furthermore, as businesses rely more on software, many develop a computer service department to assist the employees with product information, software installation and solutions to software problems. To assist in this area, computer service departments will install “help desks” as a mechanism to assist the employees with their software needs. Often those monitoring the help desk will encounter the same question many times, particularly if that question relates to a task involving the software program that is not explained well in the software manual, in the online help or is related to how the software is used to perform a specialized business function. Help desk administrators will often develop user manuals or distribute special bulletins, either by paper or through a webpage on the business website, providing answers to users most frequently asked questions. In addition, software vendors may provide a “Frequently Asked Questions” bulletin or webpage providing answers to typical user questions. These “typical” user questions may not be anticipated by software developers during the initial development phase and, therefore, may not be included in online help or not readily accessible in the user manuals.
The result of this activity in the software help area is that the user is faced with many places to check in the event a question arises regarding a specific function of the software. In addition, help desk administrators are forced to maintain many sources of help information so that when a user calls in with a question, the help administrator is faced with checking various help bulletins, “Frequently Asked Questions” lists and online help in order to address the users questions.
Unfortunately, it is impossible for software developers to anticipate all the desired help related information to include in a help file during the software development phase. Regardless of the type of search engine used to support the help function, prior search engines typically do not allow a user to include additional information to be efficiently searched outside of what is in the help files. Prior search engines do not allow a user to include specific queries to relate to specific help topics. To address this help source overload by the user, an alternative approach allows the user to add a file as a help file for a software program. However, this approach does not provide for the addition of user queries to associate with specific help topics.
Therefore, there exists a need to search files that do not have a format supported by a search engine. In addition, there exists a need to associate user queries with related help topics contained in a file that is formatted in a manner supported by a search engine.
SUMMARY OF THE INVENTION
The present invention meets the needs described above by providing a method for searching a file having a file format not supported by a search engine. Advantageously, the present invention can search this file in combination with other help files, and associate user queries with related help topics contained in the unsupported file.
Generally described, the present invention initiates the file search by retrieving a list of files from a user. The list of files typically comprises compressed HTML files or paths and their associated addresses to webpages. The compressed HTML file or webpage each contains help titles, which identify the topic of help information, referred to as title data. In addition, the compressed HTML file or webpage each also contains the help information identified by the help title, referred to as body data. Each file is parsed to retrieve the title data and body data. Using the title tags and body tags located within the file, the data associated with the title and body tags is retrieved. The body data is related to its corresponding title data. In addition, user queries can be received so that the user may associate the query with the title data.
Term-topic links can be created by using the retrieved data. Terms are identified from the title data, related body data and user queries, whereas a topic is determined using the title data. The term-topic links are then created by linking the identified terms to the topic. Term-topic lists are generated for those terms that are linked to multiple topics.
Following the creation of the term-topic links, heuristics are used to assign a probability to each of the term-topic links. For one aspect of the present invention, the heuristics define nouns to have a higher probability than verbs, verbs to have a higher probability than adjectives and adjectives and adverbs to have the same probability. In the event it can't be determined whether the term is a noun, verb, adjective or adverb, the term is assigned a probability equal to that of a noun.
After the application of heuristics, the present invention can train the term-topic links based on user-entered queries and associated target topics. This allows the user to indirectly modify the probabilities associated with the term-topic links by requiring a target topic to be retrieved with a high combined probability. Using a query, a list of topics is retrieved based on the terms found in the query. The combined probabilities of all the retrieved topics are determined. If the target topic is not contained within a predefined number of topics having the highest c
Blewett William John
Gerlt David C.
Hawley Eric
Moore Charles G.
Homere Jean R.
Kilpatrick & Stockton LLP
Microsoft Corporation
Pratt, Esq John S.
LandOfFree
Method for searching a file having a format unsupported by a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for searching a file having a format unsupported by a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for searching a file having a format unsupported by a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2558922