Conducting the hierarchical cluster analysis, in fuzzy theory, we often use transitive closure (by Zadeh) to make a matrix follow the transitive law. However, it tends to make the entries in the matrix greatly different from the entries in the original data. With this tendency, various methods have been proposed in the fields of statistical analysis and multivariate analysis, etc. Here, we will not only introduce some typical methods, but also illustrate an evaluation method of their clustering results. On the other hand, the number of clusters may have to be decided in the actual cluster analysis. That is, this is a problem of which cutting level is optimal for a partition tree. Concerning this problem, while the steepest decent method in multivariate analysis and the AIC method in statistical analysis have been designed, we will propose a fuzzy decision method which is based on the evaluation function paying attention to the size and number of clusters at each level. In this paper, we further detail the practical effectiveness of our method through an application to sociometry analysis.

View full abstract

In this paper, we present optimal models to predict the survival rate of breast cancer patients in five years. Material and Methods: This study examined the 37,256 follow-up patients by 2002 that were diagnosed as breast cancer and registered in the SEER program from 1992 to 1997. We implemented seven common algorithms (Logistic Regression model, Artificial Neural Network (ANN), Naive Bayes, Bayes Net, Decision Trees with naive Bayes, Decision Trees (ID3) and Decision Trees (J48)) besides the most widely used statistical method (Logistic Regression model) to develop the prediction models. Results: The accuracy was 85.8±0.2%, 84.5±1.4%, 83.9±0.2%, 83.9±0.2%, 84.2±0.2%, 82.3±0.2%, 85.6±0.2% for the Logistic Regression model, ANN, Naive Bayes, Bayes Net, Decision Trees with naive Bayes, ID3 and J48, respectively. Conclusion: In this study, Logistic Regression model showed the highest accuracy. The J48 had the highest sensitivity and the ANN had the highest specificity. The Decision Trees models tend to show high sensitivity. And the Bayesian models were apt to show the accuracy going up. We found that the optimal algorithm might be different by the predicted objects and dataset.

View full abstract

This paper describes a new concept for the nervous processing mechanism based on Cosmos Diagram. The idea on the Cosmos Diagram has been emerged in 1989, and later, it is applied to create the fundamental theory for the artificial emotional system based on Fuzzy theory and the subjective observation, and further, the applicability to the expression of dynamical activity of nervous system has just been presented at the international BMFSA conference. At first, the general theory of the Cosmos Diagram is introduced, and the next, it is shown that the application of Cosmos Diagram is very effective to express the system dynamical activity, especially neural network systems.

View full abstract

We propose a new associative memory to improve its noise tolerance and storage capacity. Our underlying model is an improved multidirectional associative memory (IMAM), which uses autoassociative bottleneck neural networks to remove noise in its input, i.e., analyze patterns. IMAM has inefficient storage capacity and low noise tolerance due to a correlation matrix representing association. One of our basic ideas is to replace a correlation matrix with a multilayer perceptron (MLP), which has better learning and generalization capability. Moreover, we introduce two improvements. One is to add intermediate elements into MLP to improve its performance. The other is to use outputs of hidden layers in a five-layer bottleneck neural network. These outputs include information on synthesis of a key pattern from compressed information in the middle layer. To evaluate the proposed approaches, we compared three types of associative memory: associative memory with a bottleneck neural network and MLP (AM/B-M), AM/B-M with intermediate elements (AM/B-I), and AM/B-I with synthetic outputs (AM/B-IS). 10-by-10 images of Latin alphabet are used as patterns for association. In a case of association between 78 non-injective pattern pairs with 10% noise, our proposed AM/B-IS is better than AM/B-M by more than 40% in pattern recalling ratio.

View full abstract

A constrained optimization problem or a constraint satisfaction problem in which the constraint is a conjunction of disjunctions of linear inequalities is widely used in many fields, e.g. spatial layout and path planning of robots and vehicles. These problems are called DLP or DLCSP, respectively. In this paper we propose a neural network for the DLCSP. For a DLP/DLCSP, when a disjunct is chosen for each disjunction, the whole constraint becomes a conjunctive constraint, i.e. a linear programming (LP) or linear constraint satisfaction problem (LCSP), respectively. Hence algorithms for LP/LCSP can work as an underlying algorithm for the DLP/DLCSP. Because the computational intractability called combinatorial explosion is caused by the existence of disjunctions, the most important is how to cope with the disjunctions when the number of them is large. To do this we propose a mechanism of balancing diversive and specific explorations of the state space. We have already proposed a neural network called LPPH for the satisfiability problem (SAT). The LPPH does not trap by any point which is not a solution of the SAT, and when it comes near a solution, it converges to the solution. The proposed neural network LPPH-DLCSP inherits these properties from the LPPH. In this paper we also provide experimental results which show the effectiveness of the proposed mechanism of balancing diversive and specific explorations.

View full abstract

Quality is an attribute which is not only important for products, but also related to management and enterprise itself The development and promotion of personnel resources are indispensable in improving the quality of management and a company. The management quality relates deeply to a corporate culture and the social sense of responsibility. Therefore, companies should emphasize on personnel resource development and training for employees. In this paper, engineers of a regional IT company are inquired through questionnaires and observed talents are analyzed. Through the questionnaires, this research tries to evaluate the values to be promoted. An attempt is made to clarify the relation between QWL (Quality of Working Life) and personnel training. The paper illustrates that the management quality and CSR (Corporate Social Responsibility) of regional companies relate deeply to the improvement of their individual growth and the quality of their own.

View full abstract

The aim of this study is to build a prototype of a multi-camera tracking system that enables us to track multiple human motions at one time. Generally, to recognize a single human motion is easier than to link multiple human motions. This is because multiple human motions move to different directions, while a single human motion moves one direction in a certain period. Therefore, we need a reliable system to track multiple human motions in every frame. In this paper, we propose a new method for human tracking systems using foot step direction with prediction formula to forecast the direction of human movement. This process enables the system to predict the moving direction of an object, and also gives the signal to other cameras to aware the appearance of an object. The detail outcome and expected result are discussed in this paper.

View full abstract

Determination of logistic problem is one of major research in a supply chain network. The decision of a strategic location (in terms of cost and profits) is urgently essentials in order to make a company more competitive then other competitors in an industry. Many factors should be taken into consideration while selecting a location to build a logistic place in order to serve other retailers city effectively. There are many algorithms based on several different approaches have been proposed previously. In this paper, approaching from biologically inspired computation, a DNA computing algorithm is proposed to solve the problem. In our study, we employed cluster-based technique to group all cities before choosing a distribution center. An example is presented to illustrate how the proposed technique works.

View full abstract

Computer understanding of human social interactions is a challenging topic in the field of human computing due to its multi-party dynamic nature and multimodal characteristics. Context plays an essential role in the understanding of human behaviors during group interactions. This paper presents a novel Event Based Dynamic Context Model to represent hierarchical interaction context and solve the problems of context awareness. Sensing of dynamic context is based on multi-level event detection. Online analysis of multi-level events can be achieved by our model, which is superior over previous works. Implementations in our smart meeting room demonstrate the effectiveness of our approach.

View full abstract

We propose a model to describe cluster productions within the framework of the intranuclear cascade (INC) model. In our model, combination of the "surface coalescence" and the "knockout" is implemented to describe cluster productions in intermediate energy nuclear reactions. In the present work, the basic ingredients of our INC model are defined and applied to nucleon and cluster productions for the proton-nucleus reactions. Our INC model gives a good overall agreement with experimental data, not only double-differential cross sections of nucleon productions but also cluster productions at intermediate energies. Influence of introduction of the cluster production mechanisms is investigated.

View full abstract

We propose a new nuclear reaction simulation, termed the stochastic quantum cascade model, which deals with proton-nucleus elastic scattering and multiple inelastic reactions, for the purpose of more precise dose estimation in particle cancer therapy. Based on stochastic quantization theory, we describe the wave motion of protons using particle motion, i.e., Newtonian mechanics, and then incorporate elastic scattering into the framework of the intranuclear cascade model. Numerical calculations are carried out and compared with experimental angular distributions of cross sections. Reasonable agreements are demonstrated which verifies the high predictive ability and usefulness of the model.

View full abstract

Recently metabolic syndrome has become a major concern to health. Many of the youth have fallen victim of these syndrome or related consequences. Self-organizing maps (SOM) can be viewed as a visualization tool that performs a sort of compromise between a high-dimensional dataset and a two-dimensional plane. This paper gives an overview, analysis and visualization of the syndrome using SOM with a view to understanding its trends and inter-relating other risk factors that have no relation with the syndrome yet having a very big contribution to the syndrome and hence overall health risk.

View full abstract

Diabetes mellitus, hypertension, hyperlipidemia, and hyperuricemia are called lifestyle related disease. Gene Expression programming might be one of the best programming methods to reconstruct disease tissue histology. The arteriosclerosis represents the terminal state of vascular degeneration depending on long never-ending diabetes, hypertension, and hyperlipidemia. Histological reconstruction might induce the better disease situation. We created 5 basic commands to control vascular atherosclerosis. One: β-oxidation, 2: Activate PPP, and 3: Inhibition of collagen synthesis, those are introduced from Biochemical Expression Programming (BEP). Four: digest collagen is introduced from Gene expression Programming (GEP). Five: Suspend embolism is an example of medicine-expression programming (MEP). We succeeded to decrease the hypotensives by GEP programmed to clean up atherosclerotic blood vessels.

View full abstract