Abstract: A data extracting unit extracts from a text fact data stipulated by a combination of a target object, an attribute name, and an attribute value. A data aggregating unit groups data similar to the extracted fact data throughout the text, and aggregates the number of occurrences. An inconsistency detecting unit detects an inconsistent data group which cannot be consistent by scanning a data set aggregated by the data aggregating unit. A correctness/incorrectness determining unit determines which data is correct within the inconsistent data group. A final data integrating unit integrates and outputs correct data. Additionally, the degree of reliability is assigned to fact data when the fact data is extracted from a text, and also a correctness/incorrectness determination of each data within a data group can be made by using the degree of reliability assigned to the fact data.

Abstract: A document image search apparatus generates a text by performing the character recognition of a document image and determines a re-process scope. Then, the apparatus generates a candidate character lattice from the re-recognition result of the re-process scope, generates character strings from the candidate character lattice and adds the character strings to the text. Then, the apparatus performs index search using the text with the character strings added.

Abstract: The present invention provides a file recognition apparatus and method for recognizing specific information type with respect to a web page file group collected from the Internet or stored in other storage apparatus. The file recognition apparatus of the invention comprises: a file grouping section for classifying, from a predetermined viewpoint, the file group to be recognized by file type; a file type recognition section for recognizing the type of the files according to characteristics specific to the specific information type; and a file-type-recognition correction section for correcting the recognition result of each file in consideration of the recognition precision of all files in the group. The apparatus and method of the invention can recognize various types of information, and can obtain satisfying reorganization precision.

Abstract: A multimedia object retrieval apparatus and method for retrieving multimedia objects from structured documents containing both a multimedia object and relevant explanation text. The apparatus and method parse an input structured document into a parsing result such as an HTML DOM tree; recognize a main block in the input parsing result and output a main block annotated structured document model; extract a pair of a multimedia object and corresponding explanation, and output a structured object index such as an XML format object index; and search through the structured object index to form a target object list. The apparatus and method can be applied to various kinds of structured documents, and can extract object explanations with a high precision. The apparatus and method may also identify the relationship between the object and the title of the input structured document.

Abstract: An evaluation apparatus learns the correspondence between domains and evaluation items from a Web page group in Internet, generates an evaluation set group, and generates a specified domain evaluation set by extracting evaluation items corresponding to the specified domain from the evaluation set group. Then, it evaluates a Web page to be evaluated based on the specified domain evaluation set.

Abstract: A query-and-response processing method for analyzing the intention of a query provided by a user reduces search result information to an amount manageable for the user, sorts out the result information, and presents it in an easily readable form to the user. A search request analyzer analyzes a search request provided from the user, a search criteria generator generates search criteria, then a search executor searches through a database. A query intention analyzer analyzes the intention of a query from the user, such as a query topic, and an output formatter, based on the result of the analysis, selects items to be presented to the user from the search results and determines the output format of the search results. A presentation module receives the results and presents the data to the user.

Abstract: The accuracy of retrieving or clipping documents is improved. A document to be processed is input via a document input section. Event specifying means looks up knowledge information stored in knowledge information storing means to specify the type of an event described in the input document. Attribute value extracting means extracts, from the document, attribute values of attributes relating to the specified event. Correlating means performs a process of correlating the attribute values extracted by the attribute value extracting means with entities in the real world. Document storing means stores information (normalized information) generated by the correlating means and the document or information specifying a storage location thereof in a manner associated with each other. Document extracting means compares a query input from a user interface section with the normalized information and extracts, from the document storing means, matching documents or information specifying their storage locations.

Abstract: A hot topic extraction apparatus for extracting a hot topic from information includes an information collection unit, an information storage unit, and a hot topic extraction unit. The information collection unit collects a document from an information source. The information storage unit stores collected information. The hot topic extraction unit extracts a document from the information storage unit while the information collection unit collects the document, and retrieves a document from the retrieved document, thereby extracting a hot topic from information contained in various information sources, not extracting a hot topic from a given document.

Abstract: A document image search apparatus generates a text by performing the character recognition of a document image and determines a re-process scope. Then, the apparatus generates a candidate character lattice from the re-recognition result of the re-process scope, generates character strings from the candidate character lattice and adds the character strings to the text. Then, the apparatus performs index search using the text with the character strings added.

Abstract: An electronic news translating and delivery apparatus translates the electronic news received via a network successively, and delivers the translated news to subscribers via another network. Specifically, the apparatus comprises a user manager including a register means to register news groups reserved by user, a translator which translates the articles in the news group registered in the register means of the user manager into a suitable form for the user, a translation manager to manage the translating of the news by the translator, and a delivery manager to deliver the news translated by the translator to registered users.

Abstract: A dictionary retrieval device is constructed by a conversion character definition form for providing group IDs for character subsets, a character-group ID conversion part for replacing a character with a group ID, an input character string conversion part for converting the input character string input from the input part to an input group ID, a dictionary conversion part for converting a word dictionary to a converted word dictionary defined by a notation group ID string, and a dictionary retrieval part for retrieving the converted word dictionary by the input group ID string. The dictionary retrieval device can retrieve a word from a dictionary which could not have been retrieved from a dictionary due to an input error, in the past, by regarding elements of character set defined by a conversion character definition form as the same element.

Abstract: An apparatus which translates a document (character string) having various kinds of typographical information, such as on a font size and on a font style, as character attributes, and reflects the typographical information attached to an original text of the document. The apparatus performs a morpheme analysis with typographical information saved, by regarding a piece of typographical information as a morpheme for a document having typographical information between characters. The apparatus judges typographical information attached to each character forming a morpheme after performing a morpheme analysis for a sentence having a piece of typographical information as a character attribute, and determines morpheme typographical information when characters forming a single morpheme carry different pieces of typographical information.

Abstract: A machine-translation/electronic-mail system comprises an electronic-mail receiving unit, an electronic-mail transmitting unit, a standard dictionary used in a machine translation process, and a private-use-word registering apparatus. The private-use-word registering apparatus comprises a word-definition recognition unit for recognizing a word-definition command for defining a private-use word used by a specified user, the word-definition command existing in an electronic mail containing a document, and a temporarily registering means for temporarily storing in a private-use word dictionary the private-use word having been extracted by the word-definition recognition unit, the private-use word dictionary being temporarily used in the translation as the private-use word dictionary.