Learn More
This paper presents a new behavior of moving multi-agents, whose agents move on cells in accordance with transition probabilities, depending on other agents over resources consisting of cells. We consider interactions or co-ordination among agents such that an agent can not move to a destination cell occupied by agents more than the agents of a current(More)
In this investigation, we propose how to combine stochastis process with rule-based techniques to recognize named entity in morphological analysis. We discuss Hidden Markov Model (HMM) for tagging English texts tentatively and we focus our attention on named entity recognition. We discuss rule-based approach over n consecutive words for rule extraction and(More)
We have developed a novel railway signal control system that works as a distributed system. It consists of a central control unit (called LC) and terminal devices (called FC) that are distributed at the railroad wayside and operate signal devices. The Internet technologies and optical LAN technologies have been used as communication methods between the LC(More)
In this work, we examine and verify a <i>topic word model </i>which says each topic can be identified by means of word distribution under same author, and by using random projection, one of the dimension reduction techniques, we show we can obtain efficient and effective processing to the model. We examine Shakespeare works and show we can identify scenes(More)
In this investigation, we propose a sophisticated approach for estimating the ages of blog authors by means of stochastic process. In this technique, we give weights on every word appeared in training data, and we extract a collection of feature words to each age. Then we examine articles on Blog based on the feature information and estimate the age by(More)
In this investigation, we propose an experimental approach for word segmentation in Japanese under domain-dependent situation. We apply Conditional Random Fields (CRF) to our issue. CRF learns several probabilistic parameters from training data with specific feature functions dependent on domains. Here we propose how to define domain specific feature(More)
Recently amount of construction about railway signalling system is steadily increasing to improve the transport services for the customer. But, at present efficiency of signal construction is very low. For example it is caused by that the period for the construction is limited to very short time at mid-night or flexibility of the signalling system is low(More)
In this investigation, we propose a sophisticated approach to identify items in several dictionaries for the purpose of integration. To identity synonymous words efficiently, we discuss IR approach based on vector space modeling to sentences in item explanations, without any knowledge of grammatical analysis or any other NLP analysis. We show the usefulness(More)
Although unified modeling language (UML) is widely used in object-oriented analysis (O0A) and design (OOD) approach, the method for checking consistency is not established yet. In UML collaboration diagram, relationship and interaction among objects are expressed declaratively without any state transitions nor attribute value changes. In this research, we(More)