Identification of evolving fuzzy rule-based models

An approach to identification of evolving fuzzy rule-based (eR) models is proposed. eR models implement a method for the noniterative update of both the rule-base structure and parameters by incremental unsupervised learning. The rule-base evolves by adding more informative rules than those that previously formed the model. In addition, existing rules can be replaced with new rules based on ranking using the informative potential of the data. In this way, the rule-base structure is inherited and updated when new informative data become available, rather than being completely retrained. The adaptive nature of these evolving rule-based models, in combination with the highly transparent and compact form of fuzzy rules, makes them a promising candidate for modeling and control of complex processes, competitive to neural networks. The approach has been tested on a benchmark problem and on an air-conditioning component modeling application using data from an installation serving a real building. The results illustrate the viability and efficiency of the approach.     

Evaluation of collaborative modeling processes for knowledge articulation and alignment

The evaluation of modeling processes for years has focused on assessing the outcome of modeling, while the process of modeling itself hardly has been subject of examination. Only in recent years, with rising interest in collaborative conceptual modeling approaches, the process of modeling has gained more attention. Different streams of research have focused on examining the sequence of model manipulations or have considered the collaborative modeling to be a negotiation process and shaped evaluation criteria accordingly. This article proposes to add the perspective of articulation and alignment of the topic of modeling to the potential foci of evaluation. Its contribution is to show, how existing research originating in theories of co-construction of knowledge can be adapted for this purpose and how the adopted perspective is complementary to those already available in related work. A comparative field study applying different evaluation methods on a single real world showcase is presented to show the practical applicability of the approach and allows to identify potential for the combination of previously unrelated analytical dimensions.

     

Active and dynamic information fusion for facial expression understanding from image sequences

This paper explores the use of multisensory information fusion technique with dynamic Bayesian networks (DBN) for modeling and understanding the temporal behaviors of facial expressions in image sequences. Our facial feature detection and tracking based on active IR illumination provides reliable visual information under variable lighting and head motion. Our approach to facial expression recognition lies in the proposed dynamic and probabilistic framework based on combining DBN with Ekman's facial action coding system (FACS) for systematically modeling the dynamic and stochastic behaviors of spontaneous facial expressions. The framework not only provides a coherent and unified hierarchical probabilistic framework to represent spatial and temporal information related to facial expressions, but also allows us to actively select the most informative visual cues from the available information sources to minimize the ambiguity in recognition. The recognition of facial expressions is accomplished by fusing not only from the current visual observations, but also from the previous visual evidences. Consequently, the recognition becomes more robust and accurate through explicitly modeling temporal behavior of facial expression. In this paper, we present the theoretical foundation underlying the proposed probabilistic and dynamic framework for facial expression modeling and understanding. Experimental results demonstrate that our approach can accurately and robustly recognize spontaneous facial expressions from an image sequence under different conditions.     

Modeling wine preferences by data mining from physicochemical properties

We propose a data mining approach to predict human wine taste preferences that is based on easily available analytical tests at the certification step. A large dataset (when compared to other studies in this domain) is considered, with white and red vinho verde samples (from Portugal). Three regression techniques were applied, under a computationally efficient procedure that performs simultaneous variable and model selection. The support vector machine achieved promising results, outperforming the multiple regression and neural network methods. Such model is useful to support the oenologist wine tasting evaluations and improve wine production. Furthermore, similar techniques can help in target marketing by modeling consumer tastes from niche markets.     

Optimization via multimodel simulation

Increasing computational power and the availability of 3D printers provide new tools for the combination of modeling and experimentation. Several simulation tools can be run independently and in parallel, e.g., long running computational fluid dynamics simulations can be accompanied by experiments with 3D printers. Furthermore, results from analytical and data-driven models can be incorporated. However, there are fundamental differences between these modeling approaches: some models, e.g., analytical models, use domain knowledge, whereas data-driven models do not require any information about the underlying processes. At the same time, data-driven models require input and output data, but analytical models do not. The optimization via multimodel simulation (OMMS) approach, which is able to combine results from these different models, is introduced in this paper. We believe that OMMS improves the robustness of the optimization, accelerates the optimization-via-simulation process, and provides a unified approach. Using cyclonic dust separators as a real-world simulation problem, the feasibility of this approach is demonstrated and a proof-of-concept is presented. Cyclones are popular devices used to filter dust from the emitted flue gasses. They are applied as pre-filters in many industrial processes including energy production and grain processing facilities. Pros and cons of this multimodel optimization approach are discussed and experiences from experiments are presented.     

Multivariate Bayesian modeling of known and unknown causes of events--an application to biosurveillance

This paper investigates Bayesian modeling of known and unknown causes of events in the context of disease-outbreak detection. We introduce a multivariate Bayesian approach that models multiple evidential features of every person in the population. This approach models and detects (1) known diseases (e.g., influenza and anthrax) by using informative prior probabilities and (2) unknown diseases (e.g., a new, highly contagious respiratory virus that has never been seen before) by using relatively non-informative prior probabilities. We report the results of simulation experiments which support that this modeling method can improve the detection of new disease outbreaks in a population. A contribution of this paper is that it introduces a multivariate Bayesian approach for jointly modeling both known and unknown causes of events. Such modeling has general applicability in domains where the space of known causes is incomplete.     

Variable selection for collaborative filtering with market basket data

The market basket data in the form of a binary user–item matrix or a binary item–user matrix can be modeled as a binary classification problem, which actually tackles collaborative filtering (CF) as well as target marketing. Effective variable selection (VS) can increase the prediction accuracy as well as identify important users or items in CF as well as target marketing. Therefore, we propose two new VS approaches: a Pearson correlation‐based approach and a forward random forests regression‐based approach, comparing the performance in a variety of experimental settings. The experimental results show that the proposed VS approaches outperform the conventional approaches in the examples. Furthermore, the experimental results are more reasonable and informative than the previous experimental results because the binary misclassification error and Top‐N accuracy for the user CF, the item CF, the user modeling, and the item modeling are all considered in this paper.     

Evaluating the generalizability of the Organizational Constraints Analysis framework: a hospital bed management case study

Cognitive Work Analysis (CWA) provides useful tools for analyzing and modeling work constraints that can inform the development of systems design requirements. However, it does not provide effective tools for analyzing and modeling organizational constraints that can inform the development of organizational design requirements. By integrating organizational theories with the CWA approach, we developed the Organizational Constraints Analysis framework, a formative approach to the analysis, modeling, and design of the organization of work. In this paper, we test the generalizability of the framework by using its two analytical templates—the Organizational Constraints model and Work Possibilities diagram—to analyze the hospital bed management work domain. The research findings suggest that the concepts, investigative probes, and notations from the analytical templates can be applied to complex work domains beyond those in which it was initially developed. We conclude with suggestions for how the Organizational Constraints Analysis framework can complement CWA methods by helping researchers and practitioners develop a broader organizational perspective on the constraints that drive how work can be done in organizations.     

A continuous variable Bayesian networks model for water quality modeling: A case study of setting nitrogen criterion for small rivers and streams in Ohio,USA

We present a continuous variable Bayesian networks modeling framework that integrates the graphical representation of a Bayesian networks model with empirical model-developing approach. Our model retains the Bayesian networks model's graphical representation of hypothesized causal connections among important variables and employs conventional statistical modeling approaches for establishing functional relationships among these variables. The modeling framework avoids discretizing continuous variables and the resulting models can be updated over time when new data are available or updated using local data to develop a site-specific model. We illustrate the modeling approach using a data for establishing nutrient criteria in streams and rivers in Ohio, U.S.A.     

A novel active learning approach for the classification of hyperspectral imagery using quasi-Newton multinomial logistic regression

In this article, a novel active learning approach is proposed for the classification of hyperspectral imagery using quasi-Newton multinomial logistic regression/Davidon, Fletcher, and Powell selective variance (MLR-DFP-SV). The proposed approach consists of two main steps: (1) a fast solution for the MLR classifier, where the logistic regressors are obtained by the use of the quasi-Newton algorithm; and (2) selection of the most informative unlabelled samples. The SV method is applied to select the most informative unlabelled samples, based on the posterior density distributions. Experiments on two real hyperspectral data sets confirmed that the proposed approach can effectively select the most informative unlabelled samples and improve the classification accuracy. Three different methods – the maximum information (MI), breaking ties (BT), and minimum error (ME) methods – were also used to obtain the most informative unlabelled samples, and it was found that the new sample selection method – SV – can select more informative samples than the BT, MI, and ME methods.     

Neural coding and decoding: communication channels and quantization.

We present a novel analytical approach for studying neural encoding. As a first step we model a neural sensory system as a communication channel. Using the method of typical sequence in this context, we show that a coding scheme is an almost bijective relation between equivalence classes of stimulus/response pairs. The analysis allows a quantitative determination of the type of information encoded in neural activity patterns and, at the same time, identification of the code with which that information is represented. Due to the high dimensionality of the sets involved, such a relation is extremely difficult to quantify. To circumvent this problem, and to use whatever limited data set is available most efficiently, we use another technique from information theory--quantization. We quantize the neural responses to a reproduction set of small finite size. Among many possible quantizations, we choose one which preserves as much of the informativeness of the original stimulus/response relation as possible, through the use of an information-based distortion function. This method allows us to study coarse but highly informative approximations of a coding scheme model, and then to refine them automatically when more data become available.     

Robust control approach for input–output linearizable nonlinear systems using high‐gain disturbance observer

This paper addresses a robust control approach for a class of input–output linearizable nonlinear systems with uncertainties and modeling errors considered as unknown inputs. As known, the exact feedback linearization method can be applied to control input–output linearizable nonlinear systems, if all the states are available and modeling errors are negligible. The mentioned two prerequisites denote important problems in the field of classical nonlinear control. The solution approach developed in this contribution is using disturbance rejection by applying feedback of the uncertainties and modeling errors estimated by a specific high‐gain disturbance observer as unknown inputs. At the same time, the nonmeasured states can be calculated from the estimation of the transformed system states. The feasibility and conditions for the application of the approach on mechanical systems are discussed. A nonlinear multi‐input multi‐output mechanical system is taken as a simulation example to illustrate the application. The results show the robustness of the control design and plausible estimations of full‐rank disturbances.Copyright © 2012 John Wiley & Sons, Ltd.     

Anatomy and Empirical Evaluation of an Adaptive Web-Based Information Filtering System

A case study in adaptive information filtering systems for the Web is presented. The described system comprises two main modules, named HUMOS and WIFS. HUMOS is a user modeling system based on stereotypes. It builds and maintains long term models of individual Internet users, representing their information needs. The user model is structured as a frame containing informative words, enhanced with semantic networks. The proposed machine learning approach for the user modeling process is based on the use of an artificial neural network for stereotype assignments. WIFS is a content-based information filtering module, capable of selecting html/text documents on computer science collected from the Web according to the interests of the user. It has been created for the very purpose of the structure of the user model utilized by HUMOS. Currently, this system acts as an adaptive interface to the Web search engine ALTA VISTA^TM. An empirical evaluation of the system has been made in experimental settings. The experiments focused on the evaluation, by means of a non-parametric statistics approach, of the added value in terms of system performance given by the user modeling component; it also focused on the evaluation of the usability and user acceptance of the system. The results of the experiments are satisfactory and support the choice of a user model-based approach to information filtering on the Web.     

Performability analysis of cloud computing centers with large numbers of servers

The ability to deliver acceptable levels of quality of service is crucial for cloud systems, and this requires performance as well as availability analysis. Existing modeling attempts mainly focus on pure performance analysis; however, the software and hardware components of cloud infrastructures may have limited reliability. In this study, analytical models are presented for performability evaluation of cloud centers. A novel approximate solution approach is introduced which allows consideration of large numbers of servers. The challenges for analytical modeling of cloud systems mentioned in the literature are considered. The analytical models and solutions, therefore, are capable of considering large numbers of facility nodes typically up to orders of hundreds or thousands, and able to incorporate various traffic loads while evaluating quality of service for cloud centers together with server availabilities. The results obtained from the analytical models are presented comparatively with the results obtained from discrete event simulations for validation.     

Distribution-based aggregation for relational learning with identifier attributes

Identifier attributes—very high-dimensional categorical attributes such as particular product ids or people's names—rarely are incorporated in statistical modeling. However, they can play an important role in relational modeling: it may be informative to have communicated with a particular set of people or to have purchased a particular set of products. A key limitation of existing relational modeling techniques is how they aggregate bags (multisets) of values from related entities. The aggregations used by existing methods are simple summaries of the distributions of features of related entities: e.g., MEAN, MODE, SUM, or COUNT. This paper's main contribution is the introduction of aggregation operators that capture more information about the value distributions, by storing meta-data about value distributions and referencing this meta-data when aggregating—for example by computing class-conditional distributional distances. Such aggregations are particularly important for aggregating values from high-dimensional categorical attributes, for which the simple aggregates provide little information. In the first half of the paper we provide general guidelines for designing aggregation operators, introduce the new aggregators in the context of the relational learning system ACORA (Automated Construction of Relational Attributes), and provide theoretical justification. We also conjecture special properties of identifier attributes, e.g., they proxy for unobserved attributes and for information deeper in the relationship network. In the second half of the paper we provide extensive empirical evidence that the distribution-based aggregators indeed do facilitate modeling with high-dimensional categorical attributes, and in support of the aforementioned conjectures. Editors: Hendrik Blockeel, David Jensen and Stefan Kramer An erratum to this article is available at .     

Conductor loss on transmission lines in monolithic microwave and millimeter-wave integrated circuits

A field-theoretical analysis of conductor loss on planar transmission lines is presented. The full-wave approach employed holds for arbitrary values of metallization thickness and skin depth. Its validity is checked by comparison to measurement data on coplanar waveguide (CPW). The results are discussed and compared with modeling approaches available so far. The considerations concentrate on the CPW case where, in contrast to the microstrip structures used in hybrid ICs, a significant influence of the fields inside the lossy conductors can be detected. Finally, conclusions regarding both millimeter-wave integrated circuit design and modeling are drawn. © 1992 John Wiley & Sons, Inc.     

Using accurate AIC-based performance models to improve the scheduling of parallel applications

Predictions based on analytical performance models can be used on efficient scheduling policies in order to select adequate resources for an optimal execution in terms of throughput and response time. However, developing accurate analytical models of parallel applications is a hard issue. The TIA (Tools for Instrumenting and Analysis) modeling framework provides an easy to use modeling method for obtaining analytical models of MPI applications. This method is based on modeling selection techniques and, in particular, on Akaike’s information criterion (AIC). In this paper, first the AIC-based performance model of the HPL benchmark is obtained using the TIA modeling framework. Then the use of this model for assessing the runtime estimation on different backfilling policies is analyzed in the GridSim simulator. The behavior of these simulations is compared with the equivalent simulations based on the theoretical model of the HPL provided by its developers.     

Feature-based and Clique-based User Models for Movie Selection: A Comparative Study

The huge amount of information available in the currently evolving world wide information infrastructure at any one time can easily overwhelm end-users. One way to address the information explosion is to use an ‘information filtering agent’ which can select information according to the interest and/or need of an end-user. However, at present few information filtering agents exist for the evolving world wide multimedia information infrastructure. In this study, we evaluate the use of feature-based approaches to user modeling with the purpose of creating a filtering agent for the video-on-demand application. We evaluate several feature and clique-based models for 10 voluntary subjects who provided ratings for the movies. Our preliminary results suggest that feature-based selection can be a useful tool to recommend movies according to the taste of the user and can be as effective as a movie rating expert. We compare our feature-based approach with a clique-based approach, which has advantages where information from other users is available. This revised version was published online in July 2006 with corrections to the Cover Date.     

Interval or moments: which carry more information?

In many practical situations, we do not have enough observations to uniquely determine the corresponding probability distribution, we only have enough observations to estimate two parameters of this distribution. In such cases, the traditional statistical approach is to estimate the mean and the standard deviation. Alternatively, we can estimate the two bounds that form the range of the corresponding variable and thus, generate an interval. Which of these two approaches should we select? A natural idea is to select the most informative approach, i.e., an approach in which we need the smallest amount of additional information (in Shannon’s sense) to obtain the full information about the situation. In this paper, we follow this idea and come up with the following conclusion: in practical situations in which a 95 % confidence level is sufficient, interval bounds are more informative; however, in situations in which we need higher confidence, the moments approach is more informative.     

Regularized Bundle-Adjustment to Model Heads from Image Sequences without Calibration Data

We address the structure-from-motion problem in the context of head modeling from video sequences for which calibration data is not available. This task is made challenging by the fact that correspondences are difficult to establish due to lack of texture and that a quasi-euclidean representation is required for realism.We have developed an approach based on regularized bundle-adjustment. It takes advantage of our rough knowledge of the head's shape, in the form of a generic face model. It allows us to recover relative head-motion and epipolar geometry accurately and consistently enough to exploit a previously-developed stereo-based approach to head modeling. In this way, complete and realistic head models can be acquired with a cheap and entirely passive sensor, such as an ordinary video camera, with minimal manual intervention.We chose to demonstrate and evaluate our technique mainly in the context of head-modeling. We do so because it is the application for which all the tools required to perform the complete reconstruction are available to us. We will, however, argue that the approach is generic and could be applied to other tasks, such as body modeling, for which generic facetized models exist.