Abstract: The sorting device comprises a distribution monitoring and cell splitting section which analyzes a distribution of a sort target data group consisting sort target data and obtains an appearance frequency of the sort target data for each sorting segment based on a sorting key. A sorting segment setting section sets sorting segments for dividing the sort target data group into segments based on the analysis by the distribution monitoring and cell splitting section. A sorting mode selection section selects a quick sorting mode if the appearance frequency is equal to or higher than a preset threshold and selects a radic sorting mode if the appearance frequency is lower than the threshold, for each sorting segment. A sorting execution section executes sorting for each sorting segment based on the selected sorting mode.

Abstract: An input section inputs a document set. A normalization section calculates a similarity as a relative value between documents, with respect to combinations of a plurality of documents, in the document set. The normalization section employs the tf·idf method. In tf·idf method, a document vector and a significance of a word included in the document is used to perform normalization to convert each similarity to an absolute value.

Abstract: The present invention provides a document clustering device which evaluates similarity between each document among documents and all the documents including the concerning document itself. And according to this evaluation result, the document clustering device divides the documents into non-hierarchical clusters. Thereby, the present invention realizes a document searching system which speedily retrieves a document satisfying a search condition from an accumulation of an enormous number of documents, and a FAQ preparing system which automatically prepares FAQ from an accumulation of an enormous number of sample questions and sample answers.

Abstract: By comprising a question sentence input unit for receiving a question sentence for retrieval, a retrieval execution unit for retrieving data from a database storing data to be retrieved and extracting data similar to the question sentence inputted by the question sentence input unit, a word contribution degree calculation unit for calculating the contribution degree of a word contributing to the extraction by the retrieval execution unit and a word contribution degree output unit for the contribution degree calculated by the word contribution degree calculation unit together with the corresponding word, the word to which importance is attached can be reported to a user.

Abstract: A question extraction unit extracts predetermined question elements such as ‘operation’, ‘state’, ‘question’, etc. from question text transmitted from a questioner by electronic mail, etc. A case retrieval unit searches a question and answering case accumulation unit using a question element extracted by the question extraction unit as a retrieval key, and outputs a question case containing a matching question element to a answer presentation unit. The answer presentation unit reads an answer case corresponding to the retrieved question case from a question and answering case accumulation unit, and presents it to the questioner. Thus, the questioner can be provided with a desired answer.

Abstract: The sorting device comprises a distribution monitoring and cell splitting section which analyzes a distribution of a sort target data group consisting sort target data and obtains an appearance frequency of the sort target data for each sorting segment based on a sorting key. A sorting segment setting section sets sorting segments for dividing the sort target data group into segments based on the analysis by the distribution monitoring and cell splitting section. A sorting mode selection section selects a quick sorting mode if the appearance frequency is equal to or higher than a preset threshold and selects a radic sorting mode if the appearance frequency is lower than the threshold, for each sorting segment. A sorting execution section executes sorting for each sorting segment based on the selected sorting mode.

Abstract: The present invention relates to an index managing unit which comprises an index storage managing section for managing content information corresponding to key information constituting an index on a storage area, a management information storage managing section for managing management information about the index on the storage area, and a management information updating section for updating the management information on the basis of record information inputted as information to be added to the index and constructed in units of a set of key information and content information. The index storage managing section is composed of an area calculating section for calculating, through an arithmetic, a free area to be set when necessary on the basis of the management information at the updating of the index and an area allocating section for allocating the storage area for the storage of the index on the basis of the calculated free area.

Abstract: A compression unit compresses original data in block units and encodes compressed data, and it generates continuation flag information expressive of the block length of the compressed data at the top part of the compressed data stored in a storage unit. An expansion unit determines the block length of the remaining data of the compressed data on the basis of the continuation flag information, fetches the remaining data from the storage unit and restores the original data.

Abstract: A retrieval apparatus compresses TRIE nodes, adds code ID labels as identifying codes thereto, and retrieves an input key corresponding to the compressed TRIE nodes. The input code is converted into an inner code for each unit code. Each unit code is collated with the label of each TRIE node. When they match, the position of the next TRIE node is determined corresponding to the pointer information stored in the corresponding position and the inner code of the next unit code.