Hisatoshi Mochizuki

The importance of sharing information on corporate LAN system has been pointed out for years, and many information sharing systems are currently up and running. However, users are not really satisfied with the current systems because 1) it requires a lot of labor to register the data to be shared, and 2) it is difficult to find the registered information efficiently because of mistakes on classifications. This paper proposes a information sharing system which has resolved these problems. This system allows users to register shared information only by dragging and dropping it into shared folders and automatically rearranges the registered information. To implement these features, we propose the methods to automatically extract and learn the keywords which used to be registered by users manually, to automatically cluster the learned keywords into categories, and to improve the search efficiency. We have evaluated our system using data of a search engine, Hi-Ho e-search, and have accomplished reproduction of 104 categories out of 125 (83.2%).

In document management systems, many compound words that are invented freely can be keyword candidates. Since the method should take many costs for processing keyword candidates, it is important to prepare an efficient method to extract keywords with information about the relation between them. This paper presents a technique for storing compound keywords with information about a relation between short based and long based compound keywords by extending Aho and Corasick (AC) string pattern matching machine for a finite number of keyword. By simulation results for 38 Japanese text files, it is shown that the proposed algorithm has performed 8.46ms/KB.

To carry out public relations activities effectively, it is essential to keep track of orientation transitions of individual customer as well as market trends. However, it is difficult to manage customer orientations using two-dimensional vectors of customer axis and oriented area axis because two dimensions are not enough to represent market trends and customer orientations that change as time passes. To solve this problem, this paper proposes a management method of customer orientations with three-dimensional vectors, which have time axis in additions to customer and oriented area axes. The information volume is rapidly increased in three-dimensional vectors. The proposed method solves this problem by storing only the positions and the values of points where information changes as time passes. In this method, only the past information is stored as positions and values, while the current situation is stored as two-dimensional vectors. This is to immediately understand the most important information, which are the current market situation and customer orientations. This proposed method enables us to conduct sales promotion considering trends in market and orientation transitions of individual customer. It also allows us to grasp the accurate effects of advertisements, which will be an index for more,effective advertisement.

An extendible hashing scheme resolves bucket overflows by reorganizing the hash function and file structure locally, so it is very suitable for fast key retrievals of dynamic key sets. However, it cannot search keys that contain a given string as substrings efficiently. In this paper, in order to design this substring search in extendible hashing, sig nature vectors are introduced as hash values, and a trie structure as an extendible hash table, where each vector is composed by a bit stream. Pseudo signature vectors are defined to identify the buckets, and a constrained depth-fir st search is presented to traverse the arcs of the trie structure. To construct a compact trie despite an increase in the number of keys, uniform signature treetops are introduced, and the method for an incremental expansion of the hash table is proposed. This approach can restrict the size of the bit stream for each key, making constrained depth-first search efficient. From simulation results, by applying the presented schemes to Japanese and English key sets, it was shown that the number of accessed buckets decreased from 40% to 10% in comparison with traditional extendible hashing for which only descriptors were used. In addition, the search time cost of the presented approach is 2-10 times faster. (C) 1998 Elsevier Science Inc. All rights reserved.

In accordance with the diffusion of applications, such as the Desk Top Publishing system, the Document Formatting system and the Document Editing system, it is easy to make a document by using a computer. However, as for allocating the diagrams(figures and tables), there are few document processing systems able to allocate diagrams on the appropriate places automatically. In a document processing system it is a very important issue to allocate diagrams on the most suitable places. This paper defines the criteria for allocating diagrams on the suitable positions by investigating published papers. These criteria concern 1) the order of diagrams to be allocated, 2) the stability of the diagram allocations, 3) the distance between the diagram and the location of the corresponding first reference in the text, 4) the allocation balance of diagrams in a text, 5) the restricted areas where diagrams shouldn't be allocated, 6) the allocation priorities between diagrams of different width. Moreover, this paper proposes a method for deciding the diagram allocations satisfying the above criteria automatically and fast on document formatting systems. In this case we have limited its application to one type of documents, which is papers. Especially, this method can skillfully allocate diagrams of different width on the page by reallocating the diagrams and texts within it, and can allocate diagrams over the document uniformly. But, this method doesn't have to reallocate diagrams, which have been already allocated, beyond the page. From the experimental results using various published papers (50 papers including 556 diagrams), 94.7 percent of the diagrams in these published papers satisfy the criteria defined in this paper, and so we confirmed the validity of the criteria. And, it has been shown that, by using this method, 99.8 percent of the diagrams satisfy the criteria. Furthermore, it has been shown that 82 percent of diagrams are allocated on the same place and 98 percent on the same page of the published paper. As for the execution time of the proposed method, it takes less than 1 second.

This paper presents a new method of LR parsing based oil the distinction of stack states and non-stack stales. Non-slack states are states which do not need to be pushed into the LR parsing stack. and slack stales are stales to be pushed into it. By using some properties based oil the stack-controlling LR parser defined here, the parsing speed and tile size of parsing tables call remarkably improved and tile improvement includes tile traditional method eliminating unit productions. By empirical observations for variety of programming languages, the efficiency is verified. Extending it to tile generalized LR parsers for natural language is also discussed.

In many applications, information retrieval is a very important research field in several key strategies, the binary trie is famous as a fast access method to be able to retrieve keys in order. However, if the binary trie is implemented, the greater the number of the registered keys, the larger storage in secondary memory is required In order to solve this problem, Jonge et al. proposed the method to change the binary trie into a compact bit stream (called the pre-order bit stream). However, searching and updating a key takes a lot of time in large key sets. This paper proposes an efficient binary digital search algorithm by introducing a new hierarchical structure. The algorithms for retrieval, insertion and deletion of keys using this new method are introduced through examples. The theoretical and experimental results, using 50,000 Japanese nouns and 50,000 English words, show that this method provides faster access than the traditional method. Retrieval is 18 similar to 20 times, the insertion is 11 similar to 13 times aid the deletion is 4 similar to 6 times faster.

A trie structure is frequently used for various applications, such as natural language dictionaries, database systems and compilers. However, the total number of states (and transitions between them) of a trie becomes large so that space cost may not be acceptable for a huge key set. In order to resolve this disadvantage, this paper presents a new scheme, called ''two-trie'', that enables us to perform efficient retrievals, insertions and deletions for the key sets. The essential idea is to construct two tries for both front and rear compressions of keys which is similar to a DAWG(Directed Acyclic Word-Graph). The approach differs from a DAWG in that the two-trie approach presented can determine uniquely information corresponding to keys while a DAWG cannot. For an efficient implementation of the two-trie, two types of data structures are introduced. The theoretical and experimental observations show that the method presented is more practical than existing ones considering the use of dynamic key sets, storing information of keys and compression of transitions.

The selection of an appropriate key search algorithm for a specific application field is an important issue in application systems development. This is because data retrieval is the most time-consuming part of many application programs. An automatic selection method for key search algorithms is presented in this paper. The methodology has been implemented in a system called KESE2 (KEy-SEarch ALgorithm SElection). Key search algorithms are selected according to the user's requirements through interaction with KESE2 which bases its inferences on an evaluation table. This evaluation table contains values rating the performance of each key search algorithm for the different searching properties, or characteristics. The selection algorithm presented is based on step by step reduction of unsuitable key search algorithms and searching properties. The paper also proposes assistance facilities that consist of both a support function and a program synthesis function. Experimental results show that the appropriate key search algorithms are effectively selected, and that the necessary number of questions asked, to select the appropriate algorithm, is reduced to less than half of the total number of possible questions. The support function is useful for the user during the selection process and the program synthesis function fully translates a selected key search algorithm into high level language in an average of less than 1 hour.

An efficient algorithm for deciding on the allocation of diagram (DAD) on automatic document layout is described. The document style is the double column setting with two columns in each page. The conditions that the allocation of the diagrams need to satisfy are highlighted. As future improvement, the allocation method for the diagram of halfway size, which is between the single and the double column diagram, should be designed.