Kyoji Kawagoe

New Time Series Data Representation ESAX for Financial Applications

Efficient and accurate similarity searching for a large amount of time series data set is an important but non-trivial problem. Many dimensionality reduction techniques have been proposed for effective representation of time series data in order to realize such similarity searching, including Singular Value Decomposition (SVD), the Discrete Fourier transform (DFT), the Adaptive Piecewise Constant Approximation (APCA), and the recently proposed Symbolic Aggregate Approximation (SAX).

A similarity search method of time series data with combination of Fourier and wavelet transforms

Time-series data, such as stock exchange rates and weather data, has widely been used in many fields. Similarity search of time-series data is important because it is useful for predicting data changes and searching for common sources. In this paper, we propose a new similarity search method of time-series data using both a discrete Fourier transform (DFT) and wavelet transform (WT). A method of reducing time-series indexing size, using a correlation coefficient, is also presented.

Toward a memory assistant companion for the individuals with mild memory impairment

With the increasing average life expectancy of world population, there are more and more dementia patients, and the needs of assistive technology emerges. According to the literature, the progress of cognitive impairment can be suppressed to be slower if the patients is constantly in calm mood. An effective way is suggested by keeping them in social relationship with others. With the goal of developing a conversational humanoid that can serve as a companion for dementia patients, we propose an autonomous virtual agent that can generate back channel feedback, such as head nods and verbal acknowledgment, on the basis of acoustic information in the users speech. The system is also capable of speech recognition and language understanding functionalities. In order to compensate the companionship of the agent and the ability to assist the users memory, we are developing a memory vest which is equipped with portable devices including an Android smartphone, two IC audio recorders, and a digital video recorder to log the daily life of the patient. The gathered activity history database can then be used to enrich the dialogue ability of the agent and for helping the user to recall his / her own memory.

Tweet-mapping Method for Tourist Spots Based on Now-Tweets and Spot-photos☆

Abstract Tourism recommender systems suggest suitable tourist spots by matching the characteristics of the tourist spots with those of the user. In this paper, we focus on an essential source of these characteristics—geotagged tweets. To solve the problem of associating geotagged tweets to tourist spots, we propose a mapping method that infers the region of a target spot on the basis of two geotagged items. The first is a geotagged tweet, which demonstrates that the tweeter was indeed at the target spot at the time the tweet was posted. We call this a “now-tweet.” The second item is a geotagged photo of the target spot, which we call a “spot-photo.” We regard these now-tweets and spot-photos as training data, and then determine the region of the tourist spot by inferring the geographical distribution of the training data. Next, we map geotagged tweets from the extracted region to the target spot. To improve the accuracy with which the tourist spot is inferred, we apply a clustering algorithm to the training data. Experimental results indicate that photo-based mapping with sophisticated training data produces the most improved performance over baseline methods. When applied to 4,559,643 geotagged tweets, our method maps them to tourist spots with an average granularity of 144.85 m.

Similarities of Frequent Following Patterns and Social Entities

Abstract Social network sites such as Twitter and Facebook are used for sharing knowledge and information among users. As social networks grow larger, it becomes difficult for a user to find frequently followed group of social entities. Recently, the frequent following pattern (FFP) mining concept and method were proposed to extract patterns of the relationship between a set of following entities and their most frequently followed entities. In this paper, we propose two similarity definitions: FFP similarity and FFP-based Entity (FbE) similarity. These similarities can be used to recommend a new appropriate social entity to a “read-only-user”. In other words, these similarities can be defined only with followed-and-following (F-F) relationships and without additional information such as entity characteristics or entity access logs. To the best of our knowledge, this is the first attempt to define these similarity definitions for social entity recommendations. Some examples show the effectiveness of our similarity definitions by checking their satisfaction of established requirement.

Music Playlist Recommendation Using Acoustic-Feature Transitions

Music is important in our daily life not only for entertainment but also for mental health. When listening to music, playlists are used to eliminate the need for individual selection. The creation of playlist is difficult and tedious for users and has been the topic of research in many studies. However, many proposed playlist generation methods are based on either similar acoustic features or meta-data similarities. In this study, we propose a new method for music playlist recommendation using acoustic feature transitions where the next song will be selected such that it naturally transitions from the current song. Our preliminary evaluations show that the proposed method is more effective compared with other methods such as random selection and nearest neighbor methods

Towards Music Information Retrieval driven by EEG signals: Architecture and preliminary experiments

Although much research on Music Information Retrieval (MIR) has been done in the last decade, the input of the current MIR to specify a user query for finding a similar piece of music is still either by the existing old-fashioned keywords or by music contents. We aim to realize a new type of MIR equipped with brain-computer interfaces using electroencephalogram (EEG) signals. Toward the new MIR, we propose an architecture of MIR driven by EEG signals in this paper. While the architecture contains many issues to be solved, the point of the architecture is to construct users music query in multi-layered aggregation of EEG signals. We describe in this paper the preliminary experiments conducted for selecting some appropriate low-level features for our multi-layered query construction and matching. It is obtained that the mental states of users while listening to music can be classified with high accuracy by using EEG signal aggregated features. We are starting development of detailed design of the architecture using the results described in the paper.

Dynamic Isolation of Network Devices Using OpenFlow for Keeping LAN Secure from Intra-LAN Attack☆

Abstract With the emergence of inexpensive network components and high-speed network services, a variety of network-capable electronic devices have become available. Typical examples of such devices include printers, network access storage (NAS), and video recorders. Because software on these devices is not always kept up-to-date, the devices are susceptible to intra-local- area-network (LAN) attacks. In this paper, a novel network system architecture is proposed to protect network devices from intra-LAN attacks by dynamically isolating infected devices with OpenFlow. Preliminary evaluation results demonstrate that the architecture is effective in actual LAN environments.

Time Series Classification Method Based on Longest Common Subsequence and Textual Approximation

Many symbolic representations of time series have been proposed by researchers over past decades. However, it is still not enough to classify time series with high accuracy in such applications as ubiquitous systems or sensor systems. In this paper, we propose a new symbolic representation of time series called l-TAX to increase the accuracy of time series classification. A time series can be represented by term sequences in l-TAX. l-TAX is based on a document like symbolic representation of time series called TAX. We use longest common subsequence as our distance measure between textually approximated time series. During time series classification, consideration of symbol sequences increases the accuracy significantly. In our evaluation, we have demonstrated that l-TAX is effective for classification as well as searching time series data set.

People and clothes recognition based on topic model

Clothes give us a lot of information. For example, when we see a person we know, his clothes are helpful to recognize him as well as his face. However, person recognition by clothes is still challenging problem in the area of computer vision due to the variety of a persons clothes, which makes his appearance highly variable. To handle this problem, we propose a person-specific appearance representation in this paper. In particular, we model a statistical generative process of images of people wearing their own clothes by topic model. From the images of people, our method can extract the appearance models of clothes appeared in it and estimate the composition of clothes which people wear for each image. The experimental results show that the proposed method is effective.