Journal of Advanced Computational Intelligence and Intelligent Informatics
Online ISSN : 1883-8014
Print ISSN : 1343-0130
ISSN-L : 1883-8014
19 巻 , 6 号
選択された号の論文の21件中1~21を表示しています
Regular Papers
  • Katsuhiro Honda, Shunnya Oshio, Akira Notsu
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 717-726
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    A close connection between fuzzy c-means (FCM) and Gaussian mixture models (GMMs) have been discussed and several extended FCM algorithms were induced by the GMMs concept, where fuzzy partitions are proved to be more useful for revealing intrinsic cluster structures than probabilistic ones. Co-clustering is a promising technique for summarizing cooccurrence information such as document-keyword frequencies. In this paper, a fuzzy co-clustering model is induced based on the multinomial mixture models (MMMs) concept, in which the degree of fuzziness of both object and item fuzzy memberships can be properly tuned. The advantages of the dual fuzzy partition are demonstrated through several experimental results including document clustering applications.

  • Daisuke Kitakoshi, Ryo Hanada, Keitarou Iwata, Masato Suzuki
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 727-737
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    This paper describes a cognitive training system to help older adults to stimulate and maintain their cognitive functions based on a memory game implemented on a tablet device. In this system, a software agent incorporated into the tablet device performs dialogic interactions with users based on the concept of human-agent interaction (HAI) to (i) reduce their psychological resistance to the system; (ii) maintain their interest in the game; and (iii) improve the motivation for users to play the game long term. The difficulty level of the game is adjusted through reinforcement learning algorithms depending on the proficiency of respective users. Several experiments and subjective evaluations by older adults were conducted to evaluate the basic characteristics of the system, and to investigate the impact of the system on cognitive function. The ultimate goal of the proposed system is to establish an environment in which users can continuously engage in dementia-prevention activities without getting bored.

  • Yuchi Kanzawa
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 738-746
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    This paper proposes three modifications for the maximizing model of spherical Bezdek-type fuzzy c-means clustering (msbFCM). First, we use multi-medoids instead of centroids (msbFMMdd), which is similar to modifying fuzzy c-means to fuzzy multi-medoids. Second, we kernelize msbFMMdd (K-msbFMMdd). msbFMMdd can only be applied to objects in the first quadrant of the unit hypersphere, whereas its kernelized form can be applied to a wider class of objects. The third modification is a spectral clustering approach to K-msbFMMdd using a certain assumption. This approach improves the local convergence problem in the original algorithm. Numerical examples demonstrate that the proposed methods can produce good results for clusters with nonlinear borders when an adequate parameter value is selected.

  • Masahiro Inuiguchi, Keisuke Washimi
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 747-758
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    In rough set approaches, decision rules are induced from a given data set consisting of attribute values and a decision value. Induced rules are used to classify new objects, but this classification is not perfect, perhaps because the given data set does not include all possible patterns. No induced decision rules are matched totally for objects having missing patterns, and partially matched decision rules are used to estimate their classes. The classification accuracy of such an object is usually lower than that of an object totally matching decision rules. To improve the classification accuracy, we propose adding supplementary rules to the induced rules, defining the supplementary rules to improve the classification accuracy of objects only partially matching decision rules. We propose an algorithm for inducing supplementary rules, considering four classifiers consisting of supplementary rules together with originally induced rules.We investigate their performance. We also compare their classification accuracies to that of conventional classifier with originally induced rules.

  • Yukihiro Hamasuna, Yasunori Endo
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 759-765
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    This paper presents a new algorithm of sequential cluster extraction based on hard c-means and hard c-medoids clustering. Sequential cluster extraction means that the algorithm extracts ‘one cluster at a time.’ A characteristic parameter, called a noise parameter, is used in noise clustering based sequential clustering. We propose a novel sequential clustering method called new sequential clustering, extracts an arbitrary number of objects as one cluster by considering the noise parameter as a variable to be optimized. Experimental results with four data sets confirm the effectiveness of our proposal. These results also show that classification results strongly depend on parameter ν and that our proposal is applicable to the first stage in a two-stage clustering algorithm.

  • Kaoru Shimada, Takashi Hanioka
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 766-777
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    We propose a method for associative contrast rule mining from an incomplete database to find interesting differences between two incomplete datasets. The associative contrast rule is defined as follows: although an association rule “if X then Y” satisfies the given importance conditions within Database A, the same rule does not satisfy the same conditions within Database B. The proposed method extracts associative contrast rules directly without generating the frequent itemsets used in conventional rule mining methods. We developed our message using the basic evolutionary graph-based optimization basic structure and a new evolutionary strategy for rule accumulation mechanism. The method realizes association analysis between two classes of an incomplete database using the chi-square test. We evaluated the performance of the method for associative contrast rule mining from the incomplete database. Experimental results showed that our proposed method extracts associative contrast rules effectively. Evaluations of the mischief for rule measurements by missing values are demonstrated. Simulation results showed the difference between using the proposed method for an incomplete database and using the database as complete.

  • Shogo Kawaguchi, Kouki Nagamune, Yuichiro Nishizawa, Shinya Oka, Daisu ...
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 778-784
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    Anterior cruciate ligament (ACL) reconstruction is one of the treatments of ACL injuries. In the surgery, the reconstructed ligament should be properly tensioned to provide a normal ligament behavior. However, the ligament tension has been measured with an extra-articular technique in past studies, while the intra-articular ligament tension is still unknown. The purpose of this study is to compare the ligament tensions between intra- and extra-articular measurements in the ACL reconstruction. Intra-articular measurement employs a micro-force sensor designed with a width and thickness same as those of the reconstructed ligament. This study performed two experiments (i.e., sensor accuracy and cadaveric study). In the sensor accuracy experiment, the accuracy of the sensor was about 3% until an applied force of 100 N. In the cadaveric study, the results of the intra- and extra-articular measurement tensions were 13.6±3.9 N and 18.7±1.3 N (n=6), respectively. The significant difference in student t-test (p-value was 0.026) between the intra- and extra- articular measurements was observed. The bending angle and friction between the graft and bone tunnel, and the shape of the intra-articular edge of tibial bone tunnel affected the intra-articular measurement in ACL reconstruction.

  • Seiki Tokunaga, Akihiro Okushi, Sachio Saiki, Shinsuke Matsumoto, Masa ...
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 785-795
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    This paper develops ReceiptLogService Platform, which enables consumers to using their personal purchase receipts, store their receipt logs, and to use the data for various consumer services. The proposed platform consists of three components: receipt scanner, ReceiptLog DB, and ReceiptLog API. The receipt scanner digitizes daily receipts, and the ReceiptLog DB manages the scanned data. The ReceiptLog API provides the receipt log as a service. The API consists of the BasicAPI, which provides fundamental access for the receipt log, whereas the MiningAPI performs a statistical analysis of the receipt log. These APIs are published as Web services, and can used by multiple applications and services for various purposes. We also conduct an experimental evaluation with actual subjects, to confirm the usefulness of services with receipt log.

  • Hugang Han, Yuta Higaki
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 796-803
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    This paper proposes two polynomial fuzzy controllers in the context of the fuzzy polynomial model with a so-called lumped disturbance. One, called regular controller, is designed only based on the control system stability, while the other, called controller with disturbance observer, is designed on the basis of both control system stability and a disturbance observer proposed in this paper. Though both controllers are able to stabilize the control system, computer simulations conclude that the latter is better than the former from the point of view of the control performance when it comes to the lumped disturbance in the system concerned.

  • Yohei Kamiya, Hirohisa Seki
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 804-809
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    In multi-relational data mining (MRDM), there have been proposed many methods for searching for patterns that involve multiple tables (relations) from a relational database. In this paper, we consider closed pattern mining from distributed multi-relational databases (MRDBs). Since the computation of MRDM is costly compared with the conventional itemset mining, we propose some efficient methods for computing closed patterns using the techniques studied in Inductive Logic Programming (ILP) and Formal Concept Analysis (FCA). Given a set of local databases, we first compute sets of their closed patterns (concepts) using a closed pattern mining algorithm tailored to MRDM, and then generate the set of closed patterns in the global database by utilizing the merge operator. We also present some experimental results, which shows the effectiveness of the proposed methods.

  • Katsuhiro Honda, Takaya Nakano, Chi-Hyon Oh, Seiki Ubukata, Akira Nots ...
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 810-817
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    The interpretability of fuzzy co-cluster partitions were shown to be improved by introducing exclusive penalties on both object and item memberships although the conventional fuzzy co-clustering adopted exclusive natures only on object memberships. In real applications, however, fully exclusive constraints may bring inappropriate influences to some items, and partially exclusive penalties should be forced reflecting the characteristics of each item. For example, in customer-product analysis, the degree of popularity of each product may be a measure of compatibility in multiple customer groups, and exclusive penalties should be forced only to some specific products. In this paper, the conventional exclusive constraint model is further modified by forcing exclusive penalties only to some selected items, and the effects of partially exclusive partition are demonstrated from the view points of not only partition quality but also collaborative filtering applicability. In a document-keyword analysis experiment, word class is shown to be useful for exclusively selecting keywords so that the interpretability of document cluster is improved. In a collaborative filtering experiment, the recommendation capability is demonstrated to be improved by considering intrinsic differences of popularity of each product.

  • Keisuke Yamazaki
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 818-824
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    In a probabilistic approach to cluster analysis, parametric models, such as a mixture of Gaussian distributions, are often used. Since the parameter is unknown, it is necessary to estimate both the parameter and the labels of the clusters. Recently, the statistical properties of Bayesian clustering have been studied. The theoretical accuracy of the label estimation has been analyzed, and it has been found to be better than the maximum-likelihood method, which is based on the expectation-maximization algorithm. However, the effect of a prior distribution on the clustering result remains unknown. The prior distribution has the parameter, which is the hyperparameter. In the present paper, we theoretically and experimentally investigate the behavior of the optimal hyperparameter, and we propose an evaluation method for the clustering result, based on the prior optimization.

  • Takaaki Kobayashi, Takeshi Shibuya, Masahiko Morita
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 825-832
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    When applying reinforcement learning (RL) algorithms such as Q-learning to real-world applications, we must consider the influence of sensor noise. The simplest way to reduce such noise influence is to additionally use other types of sensors, but this may require more state space – and probably increase redundancy. Conventional value-function approximators used to RL in continuous state-action space do not deal appropriately with such situations. The selective desensitization neural network (SDNN) has high generalization ability and robustness against noise and redundant input. We therefore propose an SDNN-based value-function approximator for Q-learning in continuous state-action space, and evaluate its performance in terms of robustness against redundant input and sensor noise. Results show that our proposal is strongly robust against noise and redundant input and enables the agent to take better actions by using additional inputs without degrading learning efficiency. These properties are eminently advantageous in real-world applications such as in robotic systems.

  • Toshihiko Watanabe, Takeshi Kamai, Tomoki Ishimaru
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 833-842
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    The computer vision approach involves many modeling problems in preventing noise caused by sensing units such as cameras and projectors. To improve computer vision modeling performance, a robust modeling technique must be developed for essential models in the system. The RANSAC and LMedS algorithms have been widely applied in such issues, but performance deteriorates as the noise ratio increases and the calculation time for algorithms tends to increase in actual applications. In this study, we propose a new fuzzy RANSAC algorithm for homography estimation based on the reinforcement learning concept. The performance of the algorithm is evaluated by modeling synthetic data and camera homography experiments. Their results found the method to be effective in improving calculation time, model optimality, and robustness in modeling performance.

  • Sagara Sumathipala, Koichi Yamada, Muneyuki Unehara, Izumi Suzuki
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 843-851
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    Protein name identification in text is an important and challenging fundamental precursor in biomedical information processing. For example, accurate identification of protein names affects the finding of protein-protein interactions from biomedical literature. In this paper, we present an efficient protein name identification technique based on a rich set of features: orthographic, morphological as well as Proteinhood features which are introduced newly in this study. The method was evaluated on GENIA corpus with the use of different machine learning algorithms. The highest values for precision 92.1%, recall 86.5% and F-measure 89.2% were achieved on Random Forest, while reducing the training and testing time significantly. We studied and showed the impact of the Proteinhood feature in protein identification as well as the effect of tuning the parameters of the machine learning algorithm.

  • Yuchi Kanzawa
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 852-860
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    In this study, two co-clustering algorithms based on Bezdek-type fuzzification of fuzzy clustering are proposed for categorical multivariate data. The two proposed algorithms are motivated by the fact that there are only two fuzzy co-clustering methods currently available – entropy regularization and quadratic regularization – whereas there are three fuzzy clustering methods for vectorial data: entropy regularization, quadratic regularization, and Bezdek-type fuzzification. The first proposed algorithm forms the basis of the second algorithm. The first algorithm is a variant of a spherical clustering method, with the kernelization of a maximizing model of Bezdek-type fuzzy clustering with multi-medoids. By interpreting the first algorithm in this way, the second algorithm, a spectral clustering approach, is obtained. Numerical examples demonstrate that the proposed algorithms can produce satisfactory results when suitable parameter values are selected.

  • Mayuka F. Kawaguchi, Michiro Kondo
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 861-866
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    This research report treats a correspondence between implicational fragment logics and fuzzy logics from the viewpoint of their algebraic semantics. The authors introduce monotone BI-algebras by loosening the axiomatic system of BCK-algebras. We also extend the algebras of fuzzy logics with weakly-associative conjunction from the case of the real unit interval to the case of a partially ordered set. As the main result of this report, it is proved that the class of monotone BI-algebras with condition (S) coincides with the class of weakly-associative conjunctive algebras.

  • Benoît Vallade, Alexandre David, Tomoharu Nakashima
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 867-879
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    This paper proposes a concept of layered framework for adjustable artificial intelligence. Artificial intelligences are used in various areas of computer science for decision making tasks. Traditionally artificial intelligences are developed in order to be used for a specific purpose within a particular software. However, this paper stands as the first step of a research in progress whose final objective is to design an artificial intelligence adjustable to every types of problems without any modification in its source code. The present work focuses on a framework of such an artificial intelligence and is conducted in the context of video games. This framework, composed of three layers, would be re-usable for all types of game.

  • Koki Kyo, Hideo Noda
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 880-891
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    In this paper, we propose a new approach for determining the unknown quantities in Banker–Charnes–Cooper models for data envelopment analysis by developing the marginal model synthesization algorithm. In this algorithm, several marginal fractional programming models are first constructed based on a simple numeric optimization. Then, a set of synthetic Banker–Charnes–Cooper models is obtained by compounding the marginal fractional programming models. A comparison of the proposed and existing approaches in terms of computational cost and stability of results shows that the former approach has distinct advantages. We also present an application of the proposed approach for analyzing the efficiency of industries in Japanese prefectures.

  • Soichiro Yokoyama, Hiroyuki Iizuka, Masahito Yamamoto
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 892-899
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    The heuristic method we propose solves the flexible job-shop scheduling problem (FJSP) using a solution construction procedure with priority rules. FJSP is more complex than classical scheduling problems in that operations are processed on one of multiple candidate machines, one of which must be selected to get a feasible solution. The solution construction procedure with priority rules is implemented on top of the efficient existing method for solving the FJSP which consists of a genetic algorithm and a local search method. The performance of the proposed method is analyzed using various benchmark problems and it is confirmed that our proposed method outperforms the existing method on problems with particular conditions. The conditions are further investigated by applying the proposed method on newly created benchmark.

  • Naohiko Kinoshita, Yasunori Endo, Akira Sugawara
    原稿種別: Paper
    2015 年 19 巻 6 号 p. 900-906
    発行日: 2015/11/20
    公開日: 2019/07/01
    ジャーナル 認証あり

    Clustering is representative unsupervised classification. Many researchers have proposed clustering algorithms based on mathematical models – methods we call model-based clustering. Clustering techniques are very useful for determining data structures, but model-based clustering is difficult to use for analyzing data correctly because we cannot select a suitable method unless we know the data structure at least partially. The new clustering algorithm we propose introduces soft computing techniques such as fuzzy reasoning in what we call linguistic-based clustering, whose features are not incident to the data structure. We verify the method’s effectiveness through numerical examples.

feedback
Top