The aim of our research is to simulate a thinking of doctor experts in diagnosis of integrating western and eastern medicine using fuzzy logic [9]. In this paper we present a design of a such fuzzy expert system using interval computation [3]. We extend the idea of Interval Based Expert Systems in [1] for the diagnosis of Integrated Western and Eastern medicine. A general structure of the system is described. We focus more on knowledge representation and inference mechanism of the system. We show the way how to combine the different inferences of western and eastern medicine in diagnosis. The implementation of the system is briefly described. Finally, some conclusion and discussion are given.

抄録全体を表示

We propose a new type of neuron model including the simulated synaptic function where the symbolized particles are generated, emitted, stored and released towards the neurons of the next layer, as a simplified analogy of the function that real neurons and synapses do by chemical particles taking the field of the reacted particles into account as a kind of sampled database.

抄録全体を表示

This paper proposes a tine-variant fuzzy reasoning method which is realized by changing weights with time for fuzzy rules of a "fuzzy singleton-type reasoning Method" which treats fuzzy rules whose consequent part is a fuzzy singleton (that is, a real number with a weight). The time-variant fuzzy reasoning method is applied to fuzzy controls in which fuzzy control rules or their groups are changed with time by varying weights of fuzzy control rules, and good control results are obtained.

抄録全体を表示

ファジイ回帰分析を解くためには,データ数に応じて多数の拘束条件の基で線形計画法を解く必要がある.すなわち,データ数の増加に伴って拘束条件が比例的に増える.現実の多数のデータを分断するとき,ファジィ回帰分析は大規模線形計画問題を解くには多大の計算量を要した.このためファジィ回帰分析の現実の問題では,ファジィ回帰分析を効率的に解くためにこの拘束条件を減少させる必要があった.本論文ではこのような問題を解消するために,次の拘束条件数を飛躍的に減少させ,ファジィ回帰分析を効率的に解く方法を提案している.

抄録全体を表示

A new approach to canonical correlation analysis in which the Fuzzy c-Means clustering algorithm is simultaneously applied. Minimization of the within-group sum-of-squared-error and maximization of the canonical correlation coefficients yields memberships of fuzzy clusters and principal components of each cluster can be analyzed from a complex of statistical variables. The relation between ADL (activity of daily life) and care ranks of advanced aged persons is analyzed.

抄録全体を表示

A characteristics common to melody lines in ENKA (a traditional-style in Japanese popular song) is described in this paper. Feature patterns extracted from the power spectrum of melody lines in Japanese popular songs are analyzed by using fuzzy clustering. From the analysis, there exists a distinctive cluster where ENKA melodies are assembled. A rough style of the melody lines in ENKA is reasoned by analogy from the feature pattern common to the cluster. A melodic style common to ENKA is cleared when we take a view of the melody lines from long time axis in comparison with a melody length. From analyzing the feature pattern, it is recognized that the rough straight-line representing a melodic line is modulated by ups and downs fluctuations and the strong correlation between a note and front note can be expected in the melodic progression in ENKA.

抄録全体を表示

This paper describes a qualitative diagnostic method of a large-scale circulatory system model. We use Coleman's dynamic model HUMAN, which consists of 25 physiological modules, over 200 variables and over 100 parameters. In order to make definite causal relations among variables effectively, the sensitivity analysis and the structural analysis are applied to the HUMAN model. The sensitivity analysis simplifies the complex model by removing weak causal relationships, and the structural analysis visualizes hierarchical relationships among variables.

抄録全体を表示

This paper presents an endoscopic diagnostic support system of depressed type early gastric carcinoma using a fuzzy integral. Choquet's integral and Yager's fuzzy measure are employed in the system. The fuzzy measures as measures of items are detected by a genetic algorithm. The detected fuzzy measures reflected the measures of the doctor's subjectivity. The result indicates that our system using Yager's fuzzy measure has applicability for endoscopic diagnostic support.

抄録全体を表示

In this paper, a new fuzzy reasoning method is proposed based on the distances between fuzzy sets, and the characteristic of the reasoning method is discussed. Between we human fellows, the real time self-revising is carried out than utilizing past knowledge, rather mainly, when the intention is investigated to each other. The purpose of this paper is to give the real time self-revising function like the human to distance type fuzzy reasoning method which authors proposed. The proposed new fuzzy reasoning method will be called a dynamical distance type fuzzy reasoning method.

抄録全体を表示

This article shows a SD-reconstruction method of dynamic MRA images of the liver. The method is based on fuzzy if-then rules that represent expert's knowledge. The membership functions used are determined using accumulative histogram of given images. Thus, each of steps runs automatically with robustness. In the experimental result, the method is applied to three dynamic MRA data sets of the liver. Qualitative evaluation by physicians were done for reconstructed 3D images. It shows that the method segments the hepatic veins precisely, and given images are essentially for surgery planning.

抄録全体を表示

In this paper, we propose a method to segment the menisci region from Tl-weighted and T2-weighted MR images. These images are taken continuously by the same system, so we can easily registrate them by manual. Our method is composed of two steps. First, we segment the candidate region of the menisci on the basis of the position of the cartilage from T2 weighted image. Second, we segment the meniscus voxels from the candidate based on fuzzy rules from Tl-weighted and T2-weighted images. Our method can successfully segment the menisci region from MR data sets. The generated visualizations will help medical doctor to diagnose the menisci with noninvasive.

抄録全体を表示

従来の伝統的微分方程式を拡張して、その係数をファジィ数にする。その解を解析的な方法と数値解法で求める。両者の解の比較からファジィ概念が、微分方程式にもたらすものについて議論する。

抄録全体を表示

In this paper, we introduce in a new natural method of fuzzy reasoning by defining some fuzzy mathematical concepts. It is well-known that Mamdani' s method never holds the property of Monotony. Our proposal method holds the monotony property.

抄録全体を表示

Eyeground image is the only one that can see directly the arteries arid veins from the outside of body. For building the automatic diagnosis support system we need to extract the outline of the blood vessel arid measure the caliber of the blood vessel. However the eyegrourid image difficult to extract the outline with the general way for almost tinged of red. So we suggest to apply fuzzy c-means method. At the result, we can extract the outline of the blood vessel. The caliber and curve of the blood vessel can be measured arid this image processing makes possible diagnosis support by eyeground photograph.

抄録全体を表示

The features of facial expression muscles were analyzed by fuzzy theory. Human face consists of five anatomical parts; frontal, orbital, nasal, oral and auricular. In human the muscles of the auricular region are functionally weak. Membership functions were applied to analyze the various features of facial muscles.

抄録全体を表示

In this observation, We propose the system which can be evaluated the explorand's eucrasia by using the acceleration plethysmogram. We used the neural network for the system. We made the input of the system a pulse height ratio, and gave it a health screening examination result as a teacher signal. It got the result that it was effective against the assessment of the synthetic eucrasia by using the acceleration plethysmogram.

抄録全体を表示

Our purpose is to apply fuzzy decision making for anemic diagnosis. It is so difficult to distinguish benign anemia from malignant one. As leukemia and aplastic anemia, which is malignant one, has the defects of stem cell (gene)levels, the diagnosis should have essentially fuzzy inference. We investigate the method of this fuzzy inference and how to make an expert system for anemic diseases.

抄録全体を表示

This paper deals with a prognosis of patients with gastrointestinal cancer. We employ a heuristic fuzzy clustering and a fuzzy rule-base method on the activity of PyNPase in normal and cancerous tissues for the prognosis of these cancers. By the heuristic clustering based on clinical measurement data concerning differentiation of PyNPase activity between normal and tumor regions, the fuzzy diagnosis rules are designed. The prognosis of patients with cancer is done by these fuzzy diagnosis rules.

抄録全体を表示

The measuring system of blood pressures using oscillometoric methods and semiconductor pressure sensors are developed to to collect, analyze data of many students.The intranet oriented system are composed of electrical blood pressures equipment with RS-232C connector and a personal computer with web browsers,Visual Basic,relational database and Windows 98.

抄録全体を表示

The purpose of this work is to consider a method which minimizes the affects of input data error on the output of a decision making system. Theinput affects are considered to be proportional to the differential of amembership function. Such membership functions as the threshold function, the ramp function, the standard membership function and the integral of the normaldistribution function as an approximation of an error distridution function ofthe input data, ware compared. Two membership functions which minimize theinput affects are proposed : an integral of the normal distribution function having the same standard deviation to the error distribution function, and an approximation of this function, a ramp function with a width as large as 4SD of the error distribution function.

抄録全体を表示

PCR and RT-PCR primers were designed by using Fuzzy Inferential Theory. The characteristic of fuzzy inferential theory is to introduce the experience into fuzzy reasoning. The PCR, RT-PCR, and antisense primers were set to contain 42-62% of GC content. For the next steps, less homologous primers were chosen versus self-mRNA, GAPDH (glyceraldehyde 3-P-dehydrogenase), and b-actin sequence. Designed PCR primers were used to analyze gene expression levels, and molar concentrations were estimated by quantitative competitive RT-PCR assays. Antisense (As-) oligonucleotideswere designed to inhibit translation of target mRNA and to contain initiation cordon AUG region of mRNA, divide sequences from 15-19 bases upstream to downstream were scanned homology on its entire mRNA. For the control, the sense oligonucleotideswere designed by the same procedure, and mismatchedoligos were designed by each nucleotide shuffling. All the inhibitory effects of designed antisenses were confirmed by increased blood pressure of rat injecting As-nNOS to NTS (Nucleus Tractus Solitarii), decreased transmigration activity by As-MIP-lb using ATL cells, and profilingene expression suppression.

抄録全体を表示

In this paper, we propose the novel self-organizing map (SOM), in which the correlation between the elements in the input vector is considered. It is not considered that the unneccessary or redundant elements might be included in the input vector, and thus the distance between the input vector and the weight vectors can be an appropriate measure of distance. We propose the new self-organizing map algorithm, in which the distance between the input vector and the weight vectors are calculated considering correlation between the elements. The effectiveness and varidity of the proposed method is verified by applying it to classification of the facial profiles.

抄録全体を表示

I have investigated for the predicting performance of drug interactions including drug properties which influence pharmacologic actions of other drugs using layered neural networks. Drug interactions mean that the pharmacologic action is different from the case of the monotherapy in a drug for reason why interactions of something exist between some drugs. In this study, I considered the drug metabolism in liver, the induction or reduction of hepatic metabolic enzymes, the kidney function, the influence of the renal failure, the absorption in stomach and the neutralization of stomach acid as drug properties. The neural networks using this study includes two subnetworks, one of the subnetwork images drug properties which influence pharmacologic actions, the other one predicts the interaction grade including the decrease, increase or no change of pharmacologic effects recognized as drug interaction. The back propagation algorithm was used as a learning method of these neural networks. The drug interaction informations of real drugs in Japan were used as learning data. I evaluated the predicting performance using these data and neural networks, and considered for the difference in the predicting power depending on learning data counts.

抄録全体を表示

Kansei information processing has several problems as follows ; i) Kansei information includes variation among each person, ii) a relation between physical feature and mental feature is not clear, and iii) the experimental data can be used only to a specific domain. In this paper, we report relations between pictorial features and impression words with experimental results, and also consider about pictorial feature among three kinds of domains.

抄録全体を表示

The rate of Sixty-five years old and over old men to the population will exceed to twenty five percentage at 2010 in Japan. It is important to support the health of these old men. We are settling the health care network at an aging town in Japan through the health care report system using Fuzzy Logic based on patient's medical information. In this paper, we propose a new system of the health-care report and the health-care network system using Fuzzy Logic and discuss the problems to apply this health-care system to support the health of old men.

抄録全体を表示