Tomomasa Nagashima

Global bifurcation structure of chaotic neural networks and its application to traveling salesman problems

Abstract This paper studies global bifurcation structure of the chaotic neural networks applied to solve the traveling salesman problem (TSP). The bifurcation analysis clarifies the dynamical basis of the chaotic neuro-dynamics which itinerates a variety of network states associated with possible solutions of TSP and efficiently ‘searches’ for the optimum or near-optimum solutions. By following the detailed merging process of chaotic attractors via crises, we find that the crisis-induced intermittent switches among the ruins of the previous localized chaotic attractors underly the ‘chaotic search’ for TSP solutions. On the basis of the present study, efficiency of the ‘chaotic search’ to optimization problems is discussed and a guideline is provided for tuning the bifurcation parameter value which gives rise to efficient ‘chaotic search’.

Accurate and robust gene selection for disease classification using a simple statistic.

Discrimination of disease patients based on gene expression data is a crucial problem in clinical area. An important issue to solve this problem is to find a discriminative subset of genes from thousands of genes on a microarray or DNA chip. Aiming at finding informative genes for disease classification on microarray, we present a gene selection method based on the forward variable (gene) selection method (FSM) and show, using typical public microarray datasets, that our method can extract a small set of genes being crucial for discriminating different classes with a very high accuracy almost closed to perfect classification.

Knowledge-assisted recognition of cluster boundaries in gene expression data

BACKGROUND AND MOTIVATION DNA microarray technology has made it possible to determine the expression levels of thousands of genes in parallel under multiple experimental conditions. Genome-wide analyses using DNA microarrays make a great contribution to the exploration of the dynamic state of genetic networks, and further lead to the development of new disease diagnosis technologies. An important step in the analysis of gene expression data is to classify genes with similar expression patterns into the same groups. To this end, hierarchical clustering algorithms have been widely used. Major advantages of hierarchical clustering algorithms are that investigators do not need to specify the number of clusters in advance and results are presented visually in the form of a dendrogram. However, since traditional hierarchical clustering methods simply provide results on the statistical characteristics of expression data, biological interpretations of the resulting clusters are not easy, and it requires laborious tasks to unveil hidden biological processes regulated by members in the clusters. Therefore, it has been a very difficult routine for experts. OBJECTIVE Here, we propose a novel algorithm in which cluster boundaries are determined by referring to functional annotations stored in genome databases. MATERIALS AND METHODS The algorithm first performs hierarchical clustering of gene expression profiles. Then, the cluster boundaries are determined by the Variance Inflation Factor among the Gene Function Vectors, which represents distributions of gene functions in each cluster. Our algorithm automatically specifies a cutoff that leads to functionally independent agglomerations of genes on the dendrogram derived from similarities among gene expression patterns. Finally, each cluster is annotated according to dominant gene functions within the respective cluster. RESULTS AND CONCLUSIONS In this paper, we apply our algorithm to two gene expression datasets related to cell cycle and cold stress response in budding yeast Saccharomyces cerevisiae. As a result, we show that the algorithm enables us to recognize cluster boundaries characterizing fundamental biological processes such as the Early G1, Late G1, S, G2 and M phases in cell cycles, and also provides novel annotation information that has not been obtained by traditional hierarchical clustering methods. In addition, using formal cluster validity indices, high validity of our algorithm is verified by the comparison through other popular clustering algorithms, K-means, self-organizing map and AutoClass.

An overview of Kansei engineering: a proposal of Kansei informatics toward realising safety and pleasantness of individuals in information network society

Kansei engineering is the newly proposed engineering discipline having a novel and unique goal. While its aim has been considered, from the beginning, to construct methodology and technology capable of providing industrial products and services reflecting users personal preference/requirement and being evaluated by satisfaction of users, its role is now growing up to consider a new leaf of more fundamental technical issues on Kansei informatics or human-computer interactions toward realising safety and pleasantness of individuals, which will be considered to be a most important fundamental problem in the coming information network society. In this paper, we try to describe an overview of Kansei engineering referring to biometrics. The contents include the background, objectives, present status, and present technical topics of Kansei engineering, where we also introduce concrete examples of developing researches. Future developments are also discussed.

Statistical method for classifying cries of baby based on pattern recognition of power spectrum

We develop a novel method applicable to classify the causes of crying infant based on pattern recognition of power spectrum of the cry. In our frame relied on F-value, it is available in power spectrum to order a statistical significance of frequency points which will contribute to the classification of cries. We verify performance of the method by taking the painful cries of infant with the genetic disease (ADEL). The result of our method achieves an excellent prediction. We also give a discussion on the relation between the set of frequency points extracted and the formants of cries.

Introduction to Kansei Engineering

Aiming at improving the quality of life (QOL) of people in everyday living, Kansei Engineering (KE) has been proposed. Since then, while industrial sectors in Japan have applied KE, its methodology has not made much marked progress. The fact that the foundation of KE differs from any of modern technologies seems to bring some difficulties to its application. In order to reduce such difficulties, we begin with the fundamental definition of Kansei and a brief introduction to the characteristic elements of technologies in KE.

Temporal Development of Heartbeat Intervals in Transition of Sound Stimuli Inducing Different Relaxation Feelings

This paper aims to investigate the temporal development of heartbeat intervals in a transition of different sound stimuli. In the listening experiment, two minutes relaxing musical piece and white noise were employed as the different sound stimuli, and the sound stimuli were played twice alternately after five minutes rest. After the exposure of the sound stimuli, a questionnaire asked the subjects relaxation feeling for rest, noise and music sections, respectively. Psychological result showed that music section induced the highest relaxation feeling and noise section induced the lowest relaxation feelings. Average of heartbeat intervals in music section was larger than that in noise section significantly (P≪0.05). Furthermore, after the transition of sound stimuli, it took about 30-s to change heartbeat interval from previous section. These results support the previous findings that investigated the effects of sound stimuli on heartbeat and will contribute to develop musical system reflecting a users KANSEI information automatically.

LEAF: Leave-One-Out Forward Selection Method for Cancer Classification Using Gene Expression Data

Recent progress of bioinformatics technology has enabled large-scale screening of biomarker candidates. In this paper, we propose a new method called LEAF: LEAve-one-out Forward selection method for analysis of the gene expression data. Our proposed method has made it possible to construct the ranking of informative genes using the parameter which evaluates the efficiency of the class discriminant called Discriminant Power Score (DPS). We apply the LEAF to the three kinds of leukemia dataset (ALL/AML, ALL/MLL and MLL/AML), in a public database. Consequently, our method showed a stable discriminant result with 100% accuracy by the discriminant model which used the three genes set. Furthermore, it was shown that some genes with high DPS are genes related to the cancer clarified by research in recent years. In conclusion, our class discriminant method provides a high accuracy and simply result and supports discovery of a new biomarker. Our compatible method (LEAF) will be a useful tool for many researchers engaged in bioinformatics.

Experimental study on arranging music by chaotic neural network

Based on the recalling ability on dynamic (chaotic) associative memory of neural networks, we have proposed two methods for making variations of an original melody. By computer simulations, we have shown candidates for the variation of the original melody taken from the first 16 bars of Minuet G major by Bach. The results obtained in this article may suggest a possibility that chaotic neural networks can excuse such a creative task as making variations of an original melody.

Audio classification based on a closed itemset mining algorithm

Automatic audio classification is a major topic in the fields of pattern recognition and data mining. This paper describes a new rule-based classification method (cREAD: classification Rule Extraction for Audio Data) for multi-class audio data. Typically, rule-based classification requires much computation cost to find rules from large datasets because of combinatorial search problem. To achieve efficient and fast extraction of classification rules, we take advantage of a closed itemset mining algorithm that can exhaustively extract non-redundant and condensed patterns from a transaction database within a reasonable time. The notable feature of this method is that the search space of classification rules can be dramatically reduced by searching for only closed itemsets constrained by “class label item”. In this paper, we show that our method is superior to the other salient methods on the classification accuracy of a real audio dataset.