Abstract: Character strings are extracted from a seed text which is inputted as a search condition for searching prestored object documents for a relevant document. Each object document is partitioned into a plurality of blocks, and character strings are extracted from each block. Similarity of each block to the seed text is calculated by comparing the character strings extracted from the block and the character strings extracted from the seed text. Whether or not each block is relevant to the seed text is judged by comparing the calculated similarity of the block with a preset threshold value. Based on the judgment, an “inclusion degree” of each object document (including the blocks) regarding the seed text is calculated, by which object documents relevant to the seed text are outputted.

Abstract: A text mining method whereby documents (texts) can be analyzed from a wide variety of visual points.

Abstract: In a system which delivers documents which fulfill a delivery condition set by a user, when a request to change the delivery condition is entered, the system notifies the user what documents would be no longer delivered after the change so that the user can evaluate the change of the delivery condition. To be more concrete, documents which have been delivered to users are preserved and, when it is requested by a user to change the user's delivery condition, the system applies the changed delivery condition to the preserved documents and presents what documents would be no longer delivered to the user due to inconsistency with the new delivery condition.

Abstract: In a text mining technique, if the system only extracts characteristic words and phrases frequently cooccurring with the respective components of an analysis axis as an analysis condition, similar words and phrases are extracted for any component. To clearly indicate existence of characteristic words and phrases which do not appear as cooccurrence words and phrases for other components of the analysis axis, it is desired to appropriately present distinguishable features between the components to the user. For this purpose, the frequency of appearances of a plurality of characteristic words and phrases in a document satisfying each analysis condition is calculated. As a result, multiple cooccurrence words and phrases and component-cooccurrence words and phrases are discriminatively displayed. It is therefore possible for the user to appropriately analyze the contents of a plurality of documents.

Abstract: In document retrieval having the relevance feedback function to modify a searching profile for retrieval on the basis of a user's evaluation to evaluate a search result as pertinent or impertinent, recommencement of the relevance feedback returned to a desired time is permitted. An evaluation inputted by a user, a searching profile modified by the evaluation and a search result based on the searching profile are all saved while making the correspondence between them. When a request for restoration of searching profile is made, a searching profile corresponding to an evaluation designated by the user is restored.

Abstract: A document retrieval system is provided which has a document display interface which is easy to recognize the important portions even if a document retrieved by using a query expression designated by a document or a long sentence is displayed. When a text is registered, predetermined character strings and location information which are extracted from the text are stored in a location information file. A weight of each character string is calculated by a predetermined method and is stored in a weight file. In retrieving a document, predetermined character strings are extracted from a designated query expression. A similarity is calculated between the query expression and texts in the database by using the location information and the weights acquired from the location file and the weight file. In displaying the document, character strings having the high weights are extracted from the character strings used for the retrieval.

Abstract: Retrieval conditions inputted from a plurality of users are registered. According to the retrieval conditions, a retrieval is conducted for a text inputted. As a result of the retrieval, similarity of the text is calculated for each retrieval condition. The text is delivered to users of which the retrieval condition satisfies the similarity.

Abstract: Word boundary identification operations such as morpheme analysis is performed on documents to be registered, and the top positions and the end positions of words are identified. Word boundary information is obtained based on these identification results. Search indexes are created for sub-strings of a predetermined length (n-grams) extracted from the document being registered. The search index includes document identification information as well as occurrence position information which indicates that the string is located at the n-th position from the beginning of the text data, and word boundary information for an n-gram in a document.

Abstract: A method for retrieving information from a computer system including a user terminal and a storage area, includes retrieving first information from the storage area using a first search query, the first search query having a first element and a first weight that is associated with the first element. The first search query has been formulated to retrieve target information. The first information includes at least a first data block. Second information is retrieved from the storage area using a second search query. The second search query has the first element, and a second weight that is associated with the first element. The second search query is derived from a relevance feedback provided on the first data block of the first information. An end-search criterion is provided to the user terminal. The end-search criterion provides information as to whether or not to end a first retrieval procedure for the target information.

Abstract: A seed document typifying document information delivery of which a user desires from user terminals through a communication line is registered to an information delivering service system. The system extracts feature characters from the seed document and creates a user profile for each user. Acquiring document information from an information delivering party, the system extracts the feature characters from the document information and compares them with the feature characters of the user profile. The document information is transmitted to the user terminal to which the seed document having a relevance ratio exceeding a predetermined level is registered. Receiving the delivery of the document information, the user evaluates the document information as “relevant” or “irrelevant” and inputs the evaluation result to the system. The system decides a delivery service fee to each user terminal by taking the evaluation from the user into account.

Abstract: Similar document retrieving method and system for retrieving similar documents from a document database storing plural documents written in different languages with high accuracy while suppressing retrieval noise even when difference is found in the number of registered documents in dependence on the species of description languages. Statistical information concerning the registration-subjected documents is collected on a language-by-language basis upon registration thereof. Upon retrieval of documents similar to a query document, weights of words extracted from the query document are taken into account and on a language-by-language basis by referencing the statistical information.

Abstract: Retrieval conditions inputted from a plurality of users are registered. According to the retrieval conditions, a retrieval is conducted for a text inputted. As a result of the retrieval, similarity of the text is calculated for each retrieval condition. The text is delivered to users of which the retrieval condition satisfies the similarity.

Abstract: A similar document search method includes a step of extracting a characteristic word candidate as a candidate for a characteristic word from a seeds document including desired retrieval contents, a step of extracting as characteristic words of the seeds document, when the characteristic word candidate extracted by the extracting step is a compound characteristic word including a plurality of characteristic words, the compound characteristic word and constituent characteristic words included in the compound characteristic word from the characteristic word candidate, a step of calculating, according to the characteristic words extracted by the extracting step, similarity between the seeds document and a registration document, and a step of outputting as a retrieval result a result of the similarity calculated by the similarity calculating step.

Abstract: Retrieval conditions inputted from a plurality of users are registered. According to the retrieval conditions, a retrieval is conducted for a text inputted. As a result of the retrieval, similarity of the text is calculated for each retrieval condition. The text is delivered to users of which the retrieval condition satisfies the similarity.

Abstract: A registration method for structured documents includes the steps of: preparing correspondence data between a string and a string occurrence position within a structured document for each structured document, and additionally storing the correspondence data in an occurrence frequency extracting index; and preparing a list of a character, an element containing the character and a length of the element and additionally storing the list in an element length index.

Abstract: A registration method for structured documents includes the steps of: preparing correspondence data between a string and a string occurrence position within a structured document for each structured document, and additionally storing the correspondence data in an occurrence frequency extracting index; and preparing a list of a character, an element containing the character and a length of the element and additionally storing the list in an element length index.

Abstract: A method for extracting features in contents of a document without using a word dictionary and a system using the method for accurately searching for a relevant document or documents at high speed. The method includes steps of storing character strings present in a text in a text database and possibilities appearing at boundaries of words in the text in the form of an occurrence probability file, storing occurrence frequencies of the character strings in the text as an occurrence frequency file, extracting characteristic strings from a text spcified by a user with use of the occurrence probability file, and counting occurrence frequencies thereof in the user-specified text. The method calculates similarities to the user-specified text with use of the occurrence frequency file and the occurrence frequencies in the user-specified text.

Abstract: In a text mining technique, if the system only extracts characteristic words and phrases frequently cooccurring with the respective components of an analysis axis as an analysis condition, similar words and phrases are extracted for any component. To clearly indicate existence of characteristic words and phrases which do not appear as cooccurrence words and phrases for other components of the analysis axis, it is desired to appropriately present distinguishable features between the components to the user. For this purpose, the frequency of appearances of a plurality of characteristic words and phrases in a document satisfying each analysis condition is calculated. As a result, multiple cooccurrence words and phrases and component-cooccurrence words and phrases are discriminatively displayed. It is therefore possible for the user to appropriately analyze the contents of a plurality of documents.

Abstract: A document retrieval method using a computer program includes retrieving a first set of documents using a first query expression generated by the computer program. The first set of documents is provided to a user. An evaluation of the first set of documents is received from the user. The first query expression is changed to a second query expression generated by the computer program based on the evaluation.