Catálogo de publicaciones - libros

In this paper we present a novel algorithm of finding the consensus across the clustering solutions obtained in multi-scale space for time series data. Multi-scale analysis has played an important role in signal processing. The time series data can be viewed as the one dimensional signal, and the transformed time series within each scale catch specific features of the time series, thus multi-scale analysis gives us the ability of observing time series in various views. On the other hand, the global shape of the time series is preserved by transformed data within each scale. Therefore the consensus of the clusters got from all the scales access the stability of the discovered clusters. Two main contributions of this paper are: (1) We propose a time series clustering scheme based on the consensus of clustering solutions obtained from multi-scale transformed data. (2) We propose an algorithm for finding the clustering consensus in terms of the dissimilarity of meta-clusters and Bayesian theorem.

Asian options are popular path-dependent derivatives in the financial market. However, how to price them efficiently and accurately has been a longstanding research and practical problem. No known exact pricing formulas are available to price the Asian option. Although approximate pricing formulas exist, they lack accuracy guarantees. Asian options can be priced on the lattice. A lattice divides a time interval into equal-length time steps. It is known that the value computed by the lattice converges to the true option value as → ∞. Unfortunately, only subexponential-time algorithms are available if Asian options are to be priced on the lattice without approximations. Efficient approximation algorithms are available for the lattice. The best known in the literature is an ()-time approximation lattice algorithm and an ()-time approximation PDE algorithm. Our paper suggests an ()-time lattice algorithm. Our algorithm uses a novel technique based on the method of Lagrange multipliers to minimize the approximation error. Numerical results verify the accuracy and the excellent performance of our algorithm.

Double barrier options are important path-dependent derivatives in the financial market. How to price them efficiently and accurately is thus important. Until now, no simple closed-form pricing formula for double barrier options is reported. Double barrier options can be priced on a lattice that divides a certain time interval (from option initial date to maturity date) into equal-length time steps. The pricing results obtained by the lattice algorithm converge to the true option value as → ∞, and the results oscillate significantly especially when is not large enough. To obtain an accurate pricing result without suffering from price oscillation, is required to be a large number. Unfortunately, the lattice pricing algorithm runs in () time. This paper proposes a linear-time combinatorial algorithm that can generate the same pricing results as the lattice algorithm. Thus our algorithm can handle very large ’s efficiently. This algorithm uses a novel technique based on the reflection principle and the inclusion-exclusion principle. Numerical experiments are given to verify the excellent performance of our algorithm.

In city traffic flow, vehicles driven by human interact with each other with velocity fluctuations. We simulate the such city traffic and analyze it. It should be important to research the relation between traffic signals and drivers’ strategies. Then, we prepare two CML(coupled map lattice) models to treat complex traffic system with varying their characteristics, and research the effects of traffic signals. Each model has different strategies to avoid collision between vehicles. The one is the “Mixture Model” strategy which changes vehicles’ running dynamics when their gaps have gotten narrow, and the other is the “Tracking Model” strategy which decelerate vehicles’ goal velocities to follow previous one when their gaps have gotten narrow. We compare these CML models on a oneway traffic lane under the aperiodic boundary condition. “The transition point of the jam” which separates between “free flow phase” and “congestion phase” is observed. In free flow phase, major differences are not confirmed between these models. But, when it gets congested, each model takes on dissimilar complex features stem from difference of strategies. Additionally, we research effect of signal control parameters in free flow phase. We observed that whole performance is improved as cycle length is getting shorter, at that time offset can’t be very effective. When cycle length is long, whole performance is not very lower if offset is proper.

This project consists in generating a new methodology, which aims to take into account the perception of the consumer in the new products design process. This methodology will be based on KANSEI engineering. KE helps to solve current problems in design such as reducing the costs and the deadlines, increasing the variability of the answers in terms of products, and improving the response compared to the needs of the consumer and the user.

In Europe, the integration of the perception of the consumer is currently carried out in the phase of product’s validation and not in the phase of definition. The semantic differential is one of the tools mainly used to evaluate a prototype by the consumers.

It would be relevant to integrate the perception of the consumer in the earliest phases of design generation. It is the purpose of our project of Kansei Engineering. This project aims to integrate the perception of the consumer during the phases of productś definition.

After having explained the Kansei method, and released the problems, we will develop a method, which will be experimented on the base of a design study applied in the field of pair of shoes, by analyzing the correlation between the words and design elements. In the field of industrial design, researches in product semantics intend to understand how we as human beings interpret the appearance, the use and the context of a product. Taking the product as a communication media between the designer and the user, product semantics tries to explain which messages a product expresses or represents[].

The wireless handset has various restrictions such as CPU processing power and memory size. This makes it difficult to save data, provide full service and manage resource in an efficient way. In this paper, RMS(Resource Management System), which extends the processing from an internal storage to external storage space, is proposed. Overall system structure of RMS is designed and implemented in mobile device which enables to use data by on-line at run-time, and can also manage internal files efficiently. Devices using the proposed RMS will be able to apply extended processes with the ‘Mobile Space Extension’ and will be beneficial to optimal memory space through efficient internal file management.

This paper aims at studying new ways of users integration in “emotional Design” or “Kansei engineering systems”. The main goal of this study was the integration of the trend factor in design, with an early emotional evaluation of “Trend cards” produced by the designers.

After a definition of the study context, we explain the experimental protocol which was followed. It was based on a questionnaire method involving 56 French subjects and applied in the field of shoes design. The data analysis was mainly proceed by the way of a Principal Component Analysis. The expected results were centred on the emotional evaluation of the Trendcards in order to establish further design rules for a Kansei Engineering system.

In conclusion, we can recognize important semantic effects and influences which can be used as information for the implementation of the design elements data base.

This article proposes a detection method to solve the problem of determining whether the audio data has been tampered with or not. In many situations, we need to authenticate audio data as an original or some tampered version. For example, we have to verify the owner of audio data once the right is debated. By using the proposed method, we can easily authenticate whether the audio data had been altered or not and thus make the said data are more reliable. In the scheme, the audio signals are transformed by DCT, the DC and low-frequency AC are hidden in the medium-frequency AC. With the inverse transformation of DCT, we can obtain the protected audio data. To authenticate the audio signals, the protected data are transformed by DCT; then the system extracts the embedded DC and low-frequency AC from the medium-frequency AC. After the inverse DCT, we can extract the audio data and compare with the protected data. The extracted audio data will greatly differ from the protected data if these data are altered; without tampering, the two are quite similar. Also, if the protected audio data are distorted by MP3 compression, the previous situation still holds.

In the e-learning environment for self-study style, the means to solve learner’s problems such as worry and uneasiness are very limited. In such environment, even if traditional consultation service is provided, the ability to response case by case is not enough in most cases. In real society, a learner trends to consult advisors such as teachers or familiar friends, when he has a problem which he cannot solve by himself. In this research, we proposed a new agent-based study support system, which chooses an appropriate advisor instead of the learner and adjust schedule plans between both of them. To achieve this system, advisor-choosing mechanism and schedule-adjusting mechanisms are implemented.

This paper describes a support method for programming education in self-learning style and its implementation. It is difficult and important for learners to keep their motivation. In this research, methods for catching the learning data, which cope with four learning steps; reading text, check test, coding and compile, are developed. Psychological agent, which analyzes the learning data and grasps each learner’s state based on ARCS model, is also developed. According to the diagnosis by psychological agent, interface agent gives advice and hints to the learner in visual and auditory way. Implementation of prototype system, an experiment with the prototype and its results are reported. Finally, the efficiency of proposed method and further problems are discussed.