An approach for a real-world truck delivery problem (TDP) is proposed based on fuzzy neighborhood and meta-heuristics. In order to make easier the weight setting for each of the evaluation criteria presented in the evaluation function, an integrated evaluation criterion is proposed based on fuzzy neighborhood degree concept. Furthermore, with the objective to obtain a high-quality solution in short computational time, (i) a simulated annealing based method for finding a (sub-)optimal route for each vehicle; and (ii) an evolutionary computation based method for finding a (sub-)optimal schedule for a group of vehicles are proposed. The proposed method is implemented on a personal computer using C++ language, and is evaluated on real-world data from a food company in Saitama prefecture, Japan. Compared to a scheduling expert, the proposed method has resulted in 18% lower delivery cost, with 80%-90% shorter computational time. The proposed method opens new ways to bo applied to other delivery problem such as home delivery services and mail delivery problems.

View full abstract

In the field of design, the term ""harmony"" means the aesthetic value judgment with regard to mutual relations of each element in the organization of the whole one. The purpose of this paper is the inspection of harmony in Japanese architecture Chashitsu design. The main contents of the study are analyses of rhythm, composition, and the harmony based on the former results in Chashitsu design. Furthermore, we clarify the characteristics in Chashitsu design by using the law of route N in the section ""Various aspects of composition"", and using fractal dimension as the guideline of the complexity in the section ""Various aspects of rhythm"".

View full abstract

In this research, the simulation of formation and arrangement of space is carried out about architectural planning with method of cellular automata. And the possibility to the planning by cellular automata is shown. That is, the method of a simulation is shown, the separation example of space is shown, the fusion example of space is shown, and the example of the form of space is shown, and the example of a three-dimensional solid plan is shown.

View full abstract

In this paper, a support system on spatial planning of multiple dwelling houses is proposed by using methods of Artificial Life and Genetic Algorithms (GAs). As for objective buildings, multiple dwelling houses with side corridor are employed. In this system, multiple optimizations are performed by using two GAs named inner and outer GAs. In the inner GAs, dwelling units are multiplied autonomously by using methods of cellular automata and optimal forms of each dwelling house are determined in accordance with requirements of each resident. In the outer GAs, an optimal layout of all dwelling houses in a whole building are determined in accordance with requirements of both all residents and clients. Simulations are carried out by using proposed system under different assumptions and the effectiveness of proposed system is discussed and clarified. The results of simulations show that the proposed system can perform optimizations in accordance with initial input values based on requirements of residents and clients. Therefore, the proposed system is considered to be the design support system of multiple dwelling houses taken into account of requirements of residents and clients.

View full abstract

Fire in forests and fields has strong influence on natural resources, human health, weather and climate. This research considers a system of searching a fire area in forests and fields as an effective utilization of MODIS data. It is important to develop a good system which is able to extract a range, domain and distribution of fires in forests and fields from a large quantity of image database as quickly as possible and with a high accuracy. In order to achieve high performance of a system and to improve the accuracy of the analysis, we developed Searching system for Fire Area in Forests and Fields, i.e. many PC's are connected and are clustered with each purpose of the calculations to analyze the data, MOD02 and MOD09 in MODIS.

View full abstract

Tools for Chance Discovery have been used in the wide range of applications: Marketing, product designs, earthquake predictions etc. These tools visualized the maps of relations among events, to be used as scenario maps for aiding users? decisions. However, the network models in the basis of these existing tools, represented by KeyGraph, have been sometimes confusing due to the complex structure of the output graph. In this paper, the potential model has been introduced to translate the links among events in KeyGraph, into contextual relations underlying the observed events. The visual outlook of the presented tool KeyBird is like the land surface of a real island with mountains and ridges, and users easily interprets the scenes underlying the data. The experimental results implies combining KeyGraph and KeyBird is useful for aiding real business decisions.

View full abstract

A bulletin board system (BBS) equipped with KeyGraph is proposed for supporting online chance discovery process. While conventional meeting for chance discovery with KeyGraph requires participants to meet in a meeting room, the paper aims to expand the discussion space over the Internet. In order to support online chance discovery process, the BBS has functions for assisting a user in writing scenarios with visual annotation on the KeyGraph. The system also has a function for retrieving similar scenarios. A prototype BBS is implemented, and the experimental result with test subjects shows the effectiveness of the BBS for online chance discovery process.

View full abstract

Many researches support creative activities using computers. Such systems support the creation of new ideas by displaying keywords and figures. The relationships among keywords and figures emphasize information that users are most likely unaware of. However, since most systems only display a specific scene, such sequential scenes as operation histories tend to be neglected. In this paper, we propose an interface that displays animation between two preference states to comprehend preference differences between two users. Experimental results confirm that this proposed interface interpretation can be promoted by various words.

View full abstract

The aim of this paper is to discuss the development a system for the discovery of valuable new knowledge and to develop effective sales strategies based on this knowledge by utilizing massive amounts of click stream data generated by site visitors. This paper discusses and clarifies the process by which detailed consumer behavior patterns when using the site are extracted from Internet mall retail site click stream data and how these patterns can be used as a source of new ideas for creating new marketing strategies. We will also discuss our successful use of an improved version of the genome analysis system called E-BONSAI to extract and analyze special character strings related to site visitor behavior indicated by distinctive click patterns.

View full abstract

In this paper, we develop a certainty equivalent pricing formula for weather derivatives and discuss its property in the over the counter market. First, we provide a utility based approach to find the future price of weather derivatives, where the contract is assumed to be carried out between an insurance company and an industry that run a project affected by weather index, say, the average temperature. This situation is typical in the Japanese weather derivatives market, because most contracts are sold by insurance/finance companies and their price should be determined by taking asymmetric positions into account. Using an exponential utility function, it is shown that dealings may be executed at an equilibrium price with a suitable volume adjustment. Finally, we estimate the hedge effect of weather derivatives on the electricity revenue using future and put option contracts.

View full abstract

After major surgery many patients develop signs and symptoms of generalised inflammation, which is defined as Systemic Inflammatory Response Syndrome (SIRS). To examine the systemic inflammatory response syndrome in the intensive care unit (ICU) after cardiac and thoracic surgery a retrospective study was performed on 1674 selected patients admitted in the Cardiothoracic ICU of the University Hospital of Vienna. SIRS was defined according to the American College of Chest Physicians / Society of Critical Care Medicine (ACCP / SCCM) Consensus Conference. In the first phase the moment of the first occurrence of SIRS, and of severe SIRS was determined. An SIRS episode was defined as a time interval from the beginning of SIRS until the receding of the symptoms for more than 24 hours. Based on this information, an artificial neural network (ANN) was constructed to predict severe SIRS. SIRS was present in 1544 patients (92.2%), SIRS with additional signs of organ dysfunction evolving in 76.1% of the cases; the progression took less than 24 hours in 87.9% of the cases. The presence of signs of the SIRS on the first operative day is hardly suitable for the risk prediction. A significant correlation between the number of SIRS episodes and the outcome for each individual patient was found. The number of SIRS episodes could be an accurate parameter for estimating the outcome and the treatment outlay. Keywords: data mining, knowledge acquisition, neural networks, medical application.

View full abstract

In this paper, we propose a parameter extensible typeface definition method and its application to a font designing system. With this method, a user can easily define typefaces. In addition, two applications are developed for helping font definition. The former is the typeface designing tool to design the typeface, by which a user can easily make the definition graphically. The latter is the Open Type Font(OTF) creating system, which converts the typeface definition to OTF format file. The proposed method has the following four advantages. First, the cost for designing fonts is reduced from the conventional method. Second, by employing font parameters, it can express font transformations such as weight variations. Third, it can be easily defined graphically by using the tool. Fourth, a user can create OTF format file, which is standard format for font. Finally, we apply this method to a font designing system. According to the experimental results, we confirmed that the applied system excels the conventional one in respect of the quality, reflection of desired impression and the variety of the created fonts.

View full abstract

Ontologies are key elements in the Semantic Web for providing formal definitions of concepts and relationships. Such definitions are needed to have data that could be understood and reasoned upon by machines as well as humans. However, because of the possibility of having many Ontologies in the web, alignment -- which aims providing mappings across them -- is a necessary operation. Many metrics have been defined for ontology alignment. The so-called simple metrics use linguistic or structural features of Ontological concepts to create mappings. Compound metrics, on the other hand, combine some of the simple metrics to have a better results. This paper reports our new method for compound metric creation. It is based on a supervised learning approach in data mining where a training set is used to create a neural network model, performs sensitivity analysis on it to select appropriate metrics among a set of existing ones, and finally constructs a neural network model to combine the result metrics into a compound one. Empirical results of applying it on a set of Ontologies is also shown in this paper.

View full abstract

This paper proposed a research framework to discover two modules which affected the audience satisfaction based on the analysis of linguistic features of the narrators content. It comprised two research frameworks: First, linguistic feature extraction, it used three probability estimation methods to perform SAT analysis, and formed the expression ability vector for further processing. Second, Questionnaire analysis, it collected the questionnaires from audience and formed the satisfaction vector. Finally, we use Structure Equation Modeling (SEM) to probe the mutual influence between these two groups of vector and find out the most representative linguistic characteristic factors which affect audience satisfaction.

View full abstract

Computational Learning is an essential part of Intelligent System. There are number of learning approaches and concepts for gathering knowledge, for knowledge management and effective utilization of obtained knowledge. We do believe that this is a crucial part of future autonomous systems. Incremental learning, where number of Agents are gathering, sharing and utilizing information and knowledge seems to be simple but very ambitions solution for Intelligent Systems. Knowledge must be incremental, replicable and easy to read. If learning is the process of acquiring knowledge through study, experience, and teaching, how these three fashions match learning methods studied in the field of computational intelligence? Study -- memorization or reading can be projected into training of Artificial Neural Networks (ANN), which is provided off-line, using training-data file and an error function. Experience is more like a reinforcement learning concept, where knowledge is gained in interaction with environment according to the policy and prescribed value function. Teaching is the most interactive process from these three, it is mutual involvement of learner and teacher, with observation and guidance, without fixed prescribed plan. The closest match might be the Interactive Evolutionary Computation (IEC), which searches for the optimum in interaction with a human according to his or her subjective evaluation of observed results. However, IEC was probably not used in the context of learning systems yet. Ideas behind IEC might be introduced into ANN field to develop interactive learning methods for neural networks. Observation phase from IEC might be implemented using visualization techniques for neural networks, which were already developed but did not gain popularity in application areas. Selection phase from IEC, where candidates for future evaluation are chosen, can be translated into selection of perspective trends in neural network behavior. We visualize responses of particular neurons in network to its inputs, and selectively reinforce these neurons, which are favored by a human observer. This reinforcement of individual neurons can be realized by several means, amplification of outputs of neurons, bias shift, modifications of learning rate, or even reinitialization of weights. These techniques have to be further studied in order to find optimal mixture of them. We also study clustering of neurons into groups, which is necessary for visualization of practically big enough networks. Hopefully, interactive learning methods will help in the great task of computerized search of global minima, but also to bring artificial neural networks into new application areas, these which are more subjective, and more human.

View full abstract

All information processing require ...

View full abstract

Multi-hop ad-hoc networks, aka MANETs, are composed of a set of communicating mobile devices. Users of such devices are usually selfish and try to save their resources (e.g. battery life). Therefore providing services to other users is considered as costly. However at the same time there is a need to cooperate with other devices in order to achieve high level services like messaging throughout the network. We propose a game-theoretical model based on the notion of prisoner dilemma and trust in order to enable cooperation and enhance service level in such networks. Local strategies of players are optimized using a genetic algorithm. In the random-pairing prisoner dilemma, it is assumed that for each interaction, the opponent is randomly chosen. While in our model, we suppose that there will be a limited number of iterations between the opponents. This gives them the opportunity to fine-tune their strategy based on short-term memory and trust relationship. By providing a service to another user they indeed gain trust and hope that one day there will be a return, i.e. that this other user will help them back. First results indicate that it is sufficient to have very limited sets of repeated interactions in order to gain trust and cooperation. We then extend the approach to the case where authentication is provided (e.g. based on some unique IDs) and apply it the source routing. Routes are then chosen not only based on the shortest path but also on the reliability/reputation of these paths. A reputation of a path relies on the trust level that the emitter gives to intermediary nodes.

View full abstract

In this paper, we deal with the problem that will arise in the near future from robots with a lot of functions. The problem is that the robot users will have to read thick operation manuals. We designed an interaction that allows users to easily notice a robot's function without reading the manuals. We define Function Awareness as ""to notice the relationship between a user's action and a robot's action."" We propose a guideline for designing robot's actions as, ""Action Sloping,"" which allows a robot to gradually express its internal state and also allows the user to naturally notice the robot's function by observing its actions. We designed the concrete robot's actions for a sweeping robot, and the robot changes the velocity of its motion to indicate its internal state according to the distance between the robot and its user. We develop a robot that can perform Action Sloping using low-cost infrared sensors and simple rules for actions. Through experiments, we investigated the users' behaviors and clarify the problems with the proposed method.

View full abstract

The emergence of relationships between humans and artificial things in the home is analyzed from the perspective of attachment emotion. In this paper, we compare the cognitive activities of owners with attachment emotions and owners without attachment emotions; focusing in particular on the contrasts between owners of a toy doll with strong long-term attachment and short-term owners of a home robot who have not formed attachment. The method employed is protocol analysis which has been extended to the analysis of fan letters and interviews. The results indicate that attachment functions (a) to make human think more positively about things, (b) to elicit a specific action tendency, (c) to evoke the sense of subjective well-being, and (d) to enhance pro-social cognition.

View full abstract

In human-robot communications, nonverbal messages are very important. Especially, imitation is realized as an essential factor with recent developments of researches on human-robot interaction. In order to implement an imitation system using a humanoid robot, we first developed ""Robot Station,"" which is a robot-motion development environment for heterogeneous robots using abstract description. Furthermore, we are developing a real-time transformation system, which converts human's motion data from a motion capturing system into abstract descriptions in order to handle by the Robot Station. Using these developing systems, we can make heterogeneous robots imitate human's motions. In other words, robots with our systems acquire an ability to communicate nonverbal messages regardless of shapes of their bodies. This study will expand the possibilities for applications of embodiment of robots.

View full abstract

The paper proposes a new framework named Semantic Sensor Network (SS). It provides descriptions of the state of environments for physically grounded applications. The difficulty for a sensor network to describe the environment is how to extract meaningful information from sensor readings because they have little meaning by themselves. Although almost systems of sensor networks interpret the readings based on a relation between a sensory devices and an environment, it is hard for a system designer to prepare the relation one by one. SS attaches sensory devices to daily items and employs the attaching relations for the basis of describing the state of an environment. A remarkable achievement of SS is to prepare the class definitions of daily items, to generate the instance data of daily items based on the relations between sensory devices and daily items, and to infer the states of the environments based on the instance data.

View full abstract

In recent years, in the field of the entertainment robot and the nursing care robot, several methods for embedding human emotions to the robot are developed actively. We have considered a learning method of the emotion behaviors of an autonomous mobile robot by using Q-Learning method with the meta-parameters controlled by neuromodulators which are well known in the field of the brain science so far. In this paper, we propose a target selection-type Q-Learning method with the plural Q-values concerning the maximization and the minimization of rewards and punishments. We aim at the realization of a system that adaptively learns complicated emotion behaviors with the behavior selection based on the positive and negative evaluation according to the situation. Furthermore, we also report the result of an experiment by the computer simulation to confirm the efficiency of the proposed method.

View full abstract

We investigated visual information on human heads viewed from a viewpoint change under the inverted presentation by using spatial frequency (SF) analysis. Stimuli were inverted images of three familiar peoples' heads viewed from the frontal view to the back of the head and were created by four band-pass filters (8, 16, 32, and 64c/fw). Participants were required to identify each person. On the results of Experiment 1, the face inversion effect occurred in all head angles, except the back of the head. The results of Experiment 2 showed that the mean response time (RT) of the inverted frontal view and the profile view significantly decreased when the image was in high SF. In contrast, RT of the back of the head decreased on the middle range of SF. These results suggest that the different view-specific information may exist on each view, even on the inverted presentation.

View full abstract

In this paper is proposes Facial Expression Recognition system which consists of several simple probabilistic Gabor Wavelet Neural Networks(GWNN), in which the decision is made on whether the given face has the corresponding facial expression or not. Each probabilistic Gabor Wavelet Neural Network consists of feature extractor and Classifier. The feature extractor considers only 6 facial points to extract separable features, which are able to be obtained in such a way that the evaluated separability criterion is minimized by training process. In the classification process, we proposed and used probabilistic Fuzzy Neural Network Model(FNNM), which follows Bayesian minimum risk classification rule with two flexible loss coefficients. The simplified and integrated approach toward probabilistic facial expression recognition shows us a good performance and adaptation capability, and enables the system to recognize facial expressions efficiently.

View full abstract

In face recognition, learning speed of face is very important since the system should be trained again whenever a data set is added. In existing methods, training time increases rapidly with the increase of data, which leads to the difficulty of training in large dataset. Thus, we propose SVDD (Support Vector Domain Description)-based method that can learn a data set of face rapidly and reduce the computational load. And we describe the outline of useful scheme in practical use for face recognition. In experimental results, we show that the training speed of proposed method is much faster than that of other methods. Moreover, it is shown that our face recognition system can improve the accuracy gradually by learning faces incrementally from real world.

View full abstract

In this work, we suggest a method to localize DFT in spatial domain. This enables DFT algorithm to be used for local pattern matching. Once calculated, it costs same load to calculate localized DFT regardless of the size or the position of local region in spatial domain. We applied this method to face detection problem and got the results which prove the utility of our method.

View full abstract

In this paper, a method of downsizing a genetic algorithm (GA) for lips detection is described. As part of the objectives, we also try to acquire information to represent the lips. GA is very useful for optimization of parameters in complex image matching. When GA is applied to the real-time processing, the search speed is a issue with keeping the accuracy. Using small population is one of the approaches to improve the GA search speed. However, this can cause the low accuracy, because diversity of the population will be lost. In order to avoid this problem, we propose a downsized GA with automatic search domain control. In this paper, the optimization problem is template matching. The target image includes a face of talking person and has significant changes of a whole scene by camera motion. The template image is only one closed mouth and prepared for each environment and person, because of personal use. In simulations, a simple image is used, which is the basis of video sequence. Our proposed method is verified by comparison of a process of evolution between a downsized GA and a standard GA.

View full abstract

In this paper, we propose a method to design chaotic robots which have Arnold equation, Chua's equation, Hamilton equation, Lorenz equation and hyper-chaos equation. And we also propose a method to avoid obstacles which are walls and various fixed other objects. If the chaos robots were encounter with the walls and other obstacle, the robot will be avoided to obstacle by using rotational vector which have an incident angle and an angle of reflection.

View full abstract

? In this paper, we propose a method to avoid obstacles that have unstable limit cycles in a chaos trajectory surface. We assume all obstacles in the chaos trajectory surface have a Van der Pol equation with an unstable limit cycle. When a chaos robot meets an obstacle in a Lorenz equation and Hyper-chaos equation trajectory, the obstacle reflects the robot. We also show computer simulation results of Lorenz equation and Hyper-chaos equation trajectories with one or more Van der Pol as an obstacles. We proposed and verified the results of the method to make the embedding chaotic mobile robot to avoid with the chaotic trajectory in any plane. It avoids the obstacle when it meets or closes to the obstacle with dangerous degree.

View full abstract

In this paper, we propose that the synchronization method for mutual cooperative control in the chaotic mobile robot. In order to achieve the synchronization for mutual cooperative control in the chaotic mobile robot, we apply coupled synchronization technique and driven synchronization technique in the chaotic mobile robot without obstacle and with obstacle.

View full abstract

In this paper, we propose a method to target searching method that have unstable limit cycles in a chaos trajectory surface. We assume all targets in the chaos trajectory surface have a Van der Pol equation with a stable limit cycle. When a chaos robot meets the target in the Lorenz equation and Hyper-chaos equation trajectory, the target attracted the robot. We also show computer simulation results of Lorenz equation and Hyper-chaos equation trajectories with one or more Van der Pol as a target. We proposed and verified the results of the method to make the embedding chaotic mobile robot to searching target with the chaotic trajectory in any plane. It searched the target, when it meets or closes to the target.

View full abstract

This paper proposes an unconstrained handwritten Lao recognition method using alternative curvature tree. The proposed method can recognize different handwritten Lao characters by segment them into clockwise and counter clockwise handwritten segments. The main problem in using this curvature direction sequence feature is the handling of small curvature segment. Some small curvature is a real feature but some are noise or unintentional written curve. Both choices are representing in a tree format and use prototyping for recognition. The proposed method perform very well in comparing to the conventional method such as elastic matching. As the experimental results, the recognition rates was 96.78% and the recognition time (average) using the PC with 1.4 GHz Pentium 4 was 0.0204 second.

View full abstract

In this paper a method for character recognition, invariant under translation, rotation, and scaling, is addressed. The first step of the method (preprocessing) takes into account the invariant properties of the character using the axis of symmetry and a novel coding that extracts features from the character. The second step (recognition) is achieved by an Artificial Neural Network which is created by Random Weight based Cascade Correlation algorithm (RWCC), in which vectors obtained in the preprocessing step are used as inputs to it. The algorithm is tested in character recognition, using the 26 upper case letters of the alphabet. Only four different orientations and one size (for each letter) were used for training. Recognition was tested with 9 different sizes and a minimum of 36 rotations. The results are encouraging, since it achieved 98% correct recognition. Tolerance to boundary deformations and random noise was tested.

View full abstract

On-line handwriting recognition becomes important since hand-held devices become widely used. However, we still are lacking of an efficient recognition system for writer independent unconstrained Thai handwritten characters. In this paper, we propose the use of clustering algorithm using dynamic time warping as a similarity measure to analyze and categorize users's handwriting in order to construct a set of templates for each class of characters. Then, those templates are used to classify unknown handwritten sequences. To evaluate the proposed system, the experiment on the collected natural Thai handwriting were conducted. The experimental results showed the efficiency of the system.

View full abstract

This paper presents current tools and applications developed for Tablet PCs in teaching computer science and software engineering courses, presenting lectures and papers, and creating peer-review comments. The paper also proposes other Tablet PC-based handwriting applications which include crosswords, Sudoku puzzle and signature verification. Theses applications were written in Visual Studio.NET

View full abstract

An axiomatization of the Shapley value of capacities (also called fuzzy measures) is proposed. We consider capacities on general set systems of a finite universal set which are not necessarily the power sets. We shall introduce a definition of the Shapley value of fuzzy measure defined on set systems, and give a necessary and sufficient condition which characterizes our definition axiomatically. Our condition seems natural and enough understandable as axioms of the Shapley value on set systems.

View full abstract

This paper introduces a componentwise joint receipt operation on an n-component product set and constructs an additive representation for its binary relation. In the context of conjoint measurement, we present an axiom system that is adequate to the introduction of the componentwise operation. One new axiom is additive solvability, which weakens the independence axiom of n-component (n>2), additive conjoint structures. The other is multiple-indifference whose role is to give each component utility ui on Gi a similar calibration. These two and the weak ordering and weakened independence axioms lead to the additive representation.

View full abstract

Based on the greed-eficiency fairness (GEF) hypothesis and contingent focus model, we predicted that decision makers in a prisoner's dilemma game would be more likely to defect as attention to their own payoff increased and more likely to cooperate as attention to mutual payoff increased. We measured eye movement during decision making in a prisoner's dilemma game experiment; the resulting empirical data support our hypothesis.

View full abstract

Fuzzy conjoint analysis using fuzzy least squares method is proposed. The proposed conjoint analysis involves the measurement of psychological vague judgments (such as vague consumer preferences ) among choice alternatives. The proposed analysis method was a fuzzy least squares regression analysis where output data and coefficients are represented by L-R fuzzy numbers. In order to evaluate fitness for the data, a fuzzy version of a squared multiple correlation was used. An experiment was conducted to examine the effect of partial information of attribute on the overall evaluation o f desirability for tour package using the fuzzy rating method to measure vagueness in preferential judgment.

View full abstract

Because of a rapid increase of web-pages, it is getting to be difficult to retrieve a target web-page. For example, when we use a web search engine, such as Google, it often provides too many web-pages which we don't need. In order to address the issue of web search engine, we propose a new algorithm in which we can get only web-pages which we need. We use a data-mining technology to classify web pages into two subsets, the needed and not. We study two approaches the first one is decision tree learning algorithm ``ID3'' and the other is ``Document Frequency''. We propose six algorithms based on the two approaches. In this paper, we show the experimental result of our proposed algorithms in terms of recall and precision.

View full abstract

A fuzzy measure and the Choquet integral are often applied to a subjective evaluation in multi-criteria decision making. In the process of evaluation, the fuzzy measure is viewed as the weights of criteria, and the subjective evaluation is formulated by the Choquet integral with respect to the fuzzy measure. We propose a bootstrapped Choquet integral model which extends a Choquet integral model, and improves the sensitivity through the effect of interaction. The bootstrapped model reflects the criteria weights empirically according to individual data. The numerical simulations are obtained, and the behavior of bootstrapped model is compared with that of the ordinary models, in which the bootstrapped model appears to be characteristically different from others.

View full abstract

We propose an identification method of fuzzy measures using diamond pairwise comparison. Right and left side of the diamond means ordinal weights' comparison and up and down means interaction degrees. From the comparisons, we estimate a hierarchy diagram with interaction degrees and weights of evaluation items. Using the diagram, we identified the fuzzy measure using adjusted lambda fuzzy measure and Yager family.

View full abstract

An accident risk estimation method based on traffic accident data using lambda-fuzzy measure reasoning is proposed for driving safety. This method contains knowledge about the traffic situation is inferred from traffic accident data using a lambda-fuzzy measure. Experimental results confirm that the proposed method is able to predict whether an accident will happen or not, and properly estimate the risk level of the traffic situation in advance, in the same manner as expert drivers. It is applied to the assessment of the risk level in actual driving situations, and designed for implementation in driver assistant systems.

View full abstract

We investigate in this paper about the set of $k$-additive capacities dominating a given capacity, which we call the $k$-additive core. We study its structure through achievable families, which play the role of maximal chains in the classical case ($k=1$), and show that associated capacities are element (possibly a vertex) of the $k$-additive core when the capacity is $(k+1)$-monotone. The problem of finding all vertices of the $k$-additive core is still an open question.

View full abstract

The Advanced Cruse-Assist Highway Systems(AHS) are much expected to solve problems on traffic accidents. The AHS is expected not only the reducing effects of damage cost on traffic accidents but also the psychological effects like to improve comfortableness and reduce dangerous or anxious feeling. In this paper, we propose the method based on fuzzy integral to evaluate the psychological effects by introducing AHS. And we tried to construct the evaluation method by integrating fuzzy integral and conjoint analysis to execute the benefit evaluation of introducing AHS.

View full abstract

It has been already known that a set function $\mu$ defined on the power set $2^X$ of a finite set $X$ and/or the Choquet integral w.r.t. $\mu$ have a $\mathcal{C}$-decomposition iff $\mathcal{C}$ is an inclusion-exclusion covering of $X$. This paper shows that for any two positive integers $k$ and $k'$ such that $k'\le k$, there exist a finite set $X$, a $k$-monotone set function $\mu$ on $2^X$, and an inclusion-exclusion covering $\mathcal{C}$ of $X$ w.r.t. $\mu$ such that $\mu$ has no $k'$-monotone $\mathcal{C}$-decompositions.

View full abstract

A new class of fuzzy measure generated by triangular norm presented by 2 parameter is proposed. It is a distorted probability and also a generalization of lambda-fuzzy measure. Consequently, it is readily-treatable and cover wide range of fuzzy measures. We propose also the method of identification of fuzzy measures using proposed class.

View full abstract

The keyword map, which is a kind of information visualization systems, has been studied for visualizing keyword space. This system shows keywords and relationships between keywords by a graph drawing method, and is equipped with interactive features. However, existing keyword map has not considered positively the reflection of user's intention in keyword space. In this paper, the keyword map equipped with arrangement supporting functions is proposed for emphasizing user's intention. The relation between user's intentions and their usages of arrangement supporting functions is examined through experiments.

View full abstract

In this paper, we explain human support system using i-mobility. We named the intelligent car i-mobility. The first is an automatic follow system using the driver's intention recognition. The driver's intention is recognized from the speed of the car and the direction of driver's face. The second is toward joint attention system that uses driver's sight line.

View full abstract

A new self-organizing incremental network is designed for online supervised learning. During learning of the network, an adaptive similarity threshold is used to judge if new nodes are needed when online training data are introduced into the system. Nodes caused by noise are deleted to decrease the misclassification. The proposed network suits the following tasks: (1) online or even life-long supervised learning; (2) learning new information without destroying old learned information; (3) learning without any predefined optimal condition; (4) representing the topology structure of inputting online data; and (5) learning the number of nodes needed to represent every class. Experiments of artificial data and high-dimension real-world data show that the proposed method can achieve classification with a high recognition ratio, high speed, and low memory.

View full abstract