Search

Hypoglycemia is a common occurrence in critically ill patients and is associated with significant mortality and morbidity. We developed a machine learning model to predict hypoglycemia by using a multicenter intensive care unit (ICU) electronic health record dataset. Machine learning algorithms were trained and tested on patient data from the publicly available eICU Collaborative Research Database. Forty-four features including patient demographics, laboratory test results, medications, and vitals sign recordings were considered. The outcome of interest was the occurrence of a hypoglycemic event (blood glucose

The papers in this special section focus on recent developments and emerging topics in the area of deep fuzzy models that address some of the problems and limitations above. These models have been known under different names, such as hierarchical fuzzy systems and fuzzy networks. They are usually well suited for performing multiple functional compositions at either crisp or linguistic level. Deep learning has gained significant attention within the computational intelligence community in recent years. Its success has been mainly due to the increased power of modern computational platforms in terms of their ability to collect, store, and process large volumes of data. This has led to a substantial increase in the effectiveness and efficiency of data management. As a result, it has become possible to achieve high accuracy for some benchmark learning tasks, such as object classification and image recognition within a short time frame. The most common implementation of deep learning has been through neural networks due to the ability of their layers of neurons to perform multiple functional compositions as part of a multistage learning process.

In this paper, a new approach to the usage of 0-order Autonomous Learning Multi-Model (ALMMo-0) classifiers is proposed. ALMMo-0 classifiers are fully automatic and do not rely on any hyper-parameters. The creation of clouds relies on normalizing data points by their norm, which may remove an important degree of freedom from the data itself. The proposed approach consists of adding the initial radius of the clouds as an hyper-parameter, which makes it possible to skip the normalization step. This approach requires the search for the ideal value of the hyper-parameter. This way, upon training a set of models with different values for the initial radius, the user is expected to be able to choose from several models which range from more accurate to less complex. This approach was tested on three benchmark problems and compared to the results obtained using the original approach. Furthermore, this approach was also tested on a real dataset (Acute Kidney Injury). The obtained results enhance the versatility provided by the proposed method, successfully allowing the user to choose the model that fits better the design demands regarding accuracy, training time, and complexity.

As the industry continues to pursue the reduction of downtimes and increases in the efficiency of resource usage, data appears as a valuable business asset, empowering intelligent automation solutions. In order to deal with current challenges in acquisition and management of data in industrial settings, this paper proposes a framework for intelligent automation. The traditional automation pyramid does not cope with today’s application requirements, as intelligent automation needs to exchange, in realtime, information with all layers of the pyramid, simultaneously. The middleware for intelligent automation provides interoperability through integration software along with cloud functionalities and big data warehousing. The middleware also provides decentralization by connecting to remote decision support systems. This paper presents two real implementation examples of the framework in the wood and the chemical industry. The framework successfully addresses the requirements of key intelligent automation features and data characteristics, supporting the development of support decision systems into automation modules.

UID/CEC/50021/2019. Grant: FCT/MCTES TUBITAK/0008/2014 FCT/DAAD - 2018/2019 (Poc. DAAD 441.00). UID/EMS/50022/2019. project TIN2015-73566-JIN and by the Consellería de Cultura, Educación e Ordenación Universitaria (accreditation 2016–2019, ED431G/08 and reference competitive group 2019–2021, ED431C 2018/29) and the European Regional Development Fund (ERDF). The control logic underlying building automation systems has consisted, traditionally, of embedded discrete programs created using either low-level or proprietary scripting languages, or using general purpose fourth-generation visual languages like Simulink. It is also well known that programs developed in this way are hard to evolve, test, and maintain. These difficulties are intensified when continuous control problems have to be tackled or when the actuation must vary continually subject to the sensor inputs. Such is the case in day-lighting or occupancy-based control applications. In this paper, we propose a declarative high-level Domain-Specific Language that aims to reduce the effort required to specify the control logic of building automation systems. Our language combines fuzzy logic and temporal logic, enabling to define the behaviour in terms of domain abstractions. Finally, the approach has been validated in two ways: (i) in a case study that simulates the control system of an automated office room and (ii) by means of an empirical study to confirm usability (with a System Usability Scale questionnaire) and effectiveness, here regarded from the perspective of correctness, of the proposed language with respect to a well-known language like Simulink. authorsversion published

The development of dynamic data-based Decision Support Systems (DSSs) along with the increasing availability of data in the industry, makes real-time data acquisition and management a challenge. Intelligent automation appears as a holistic combination of automation with analytics and decisions made by artificial intelligence, delivering smart manufacturing and mass customization while improving resource efficiency. However, challenges towards the development of intelligent automation architectures include the lack of interoperability between systems, complex data preparation steps, and the inability to deal with both high-frequency and high-volume data in a timely fashion. This paper contributes to industrial frameworks focused on the development of standardized system architectures for Industry 4.0, closing the gap between generic architectures and physical realizations. It proposes a platform for intelligent automation relying on a gateway or middleware between field devices, enterprise databases, and DSSs in real-time scenarios. This is achieved by providing the middleware interoperability, determinism, and automatic data structuring over an industrial communication infrastructure such as the OPC UA Standard over Time Sensitive Networks (TSN). Cloud services and database warehousing used to address some of the challenges are handled using fog computing and a multi-workload database. This paper presents an implementation of the platform in the pharmaceutical industry, providing interoperability and real-time reaction capability to changes to an industrial prototype using dynamic scheduling algorithms.

Developing energy efficiency strategies for Intelligent Buildings requires a continuous monitoring of the impacts of energy efficiency measures against baseline scenarios. Since not all consumption events can be linearly described, non-linear models based on fuzzy systems and neural networks are successful where linear models either become inaccurate or too complex to be employed in practice. Herein, we study the adequacy of plain-designed non-linear models to predict energy consumption in a real-world Intelligent Building setting by using simple predictor variables such as time-of-day, weather conditions, and occupancy as estimated from WiFi traffic. The developed Fuzzy Systems and Neural Network models achieve considerably better performance and accuracy indexes than linear regression models, with VAFs equal to 79.3% and 75.7% respectively, and MAE of 5.9 and 7.1 kWh, respectively, against 36.6% VAF and 12.8 kWh MAE for linear regression models, thus demonstrating that low complexity non-linear models can be used to accurately describe energy consumption baselines in Intelligent Buildings.

Vasopressors administration in intensive care units is a risky surgical procedure that can be associated with infections, especially if done urgently such as in the case of unexpected systemic shock. The early prediction of a patient's transition to vasopressor dependence could improve overall outcomes associated with the procedure. Personalized medicine in the ICU encompasses the customization of healthcare on the level of individual patients, with diagnostic tests, monitoring interventions and treatments being fitted to the individual rather than the "average" patient. In this scope, this paper proposes an ensemble fuzzy modeling approach to a classification problem based on subgroups of patients identified by individual characteristics. A fuzzy c-means clustering algorithm was implemented to find subgroups of patients and each subgroup was used to develop a fuzzy model. The final classification of the ensemble fuzzy approach is obtained using two output selection criteria: an a priori decision criterion based on the distance from the cluster centers to the patients' characteristics, and an a posteriori decision criterion based on the uncertainty of the model output. The performance of the proposed approach is investigated using a real world clinical database and nine benchmark datasets. The ensemble fuzzy model approach performs better than the single model for the prediction of vasopressors administration in the ICU, being the a posteriori approach the best performer, with an average AUC of 0.85, showing this way the advantage of a personalized approach for patient care in the ICU.

This chapter presents two novel approaches for the identification of Takagi-Sugeno fuzzy models with time variant and time invariant features. The mixed fuzzy clustering (MFC) algorithm is used for determining the parameters of Takagi-Sugeno fuzzy models (FMs) in two different ways: (1) MFC FM, where the antecedent fuzzy sets are determined based on the partition matrix generated by the mixed fuzzy clustering algorithm; (2) FCM–UMFC FM, where the input features are transformed using MFC and the antecedent fuzzy sets are derived using fuzzy c-means (FCM). The fuzzy modeling approaches are tested on four health care applications for the classification of critically ill patients: administration of vasopressors in pancreatitis and pneumonia patients, mortality in septic shock and early readmissions. Both approaches increase the performance of Takagi-Sugeno based on FCM, in all datasets. In particular, the best performer, FCM–UMFC FM, achieves notable improvements in the four datasets.

Objective The human nucleus accumbens (Acc) has become a target for deep brain stimulation (DBS) in some neuropsychiatric disorders. Nonetheless, even with the most recent advances in neuroimaging it remains difficult to accurately delineate the Acc and closely related subcortical structures, by conventional MRI sequences. It is our purpose to perform a MRI study of the human Acc and to determine whether there are reliable anatomical landmarks that enable the precise location and identification of the nucleus and its core/shell division. Methods For the Acc identification and delineation, based on anatomical landmarks, T1WI, T1IR and STIR 3T-MR images were acquired in 10 healthy volunteers. Additionally, 32-direction DTI was obtained for Acc segmentation. Seed masks for the Acc were generated with FreeSurfer and probabilistic tractography was performed using FSL. The probability of connectivity between the seed voxels and distinct brain areas was determined and subjected to k-means clustering analysis, defining 2 different regions. Results With conventional T1WI, the Acc borders are better defined through its surrounding anatomical structures. The DTI color-coded vector maps and IR sequences add further detail in the Acc identification and delineation. Additionally, using probabilistic tractography it is possible to segment the Acc into a core and shell division and establish its structural connectivity with different brain areas. Conclusions Advanced MRI techniques allow in vivo delineation and segmentation of the human Acc and represent an additional guiding tool in the precise and safe target definition for DBS.

Fuel cell powered hybrid electric vehicles (FC-HEV) and plug-in hybrid electric vehicles (FC-PHEV) are being addressed by the automotive industry as improved and more sustainable alternative technologies relatively to conventional vehicles. Nevertheless, hybrid propulsion raises new challenges in designing the vehicle powertrain. This study highlights the significance of the driving conditions and the conflict between the optimization of investment cost, efficiency and life cycle impact (LCA) in powertrain design optimization of these kinds of vehicles. A single-objective (minimization of cost, fuel or LCA CO2eq) and multi-objective genetic algorithms (minimization of the couples cost and fuel, cost and LCA CO2eq, fuel and LCA CO2eq), linked with the vehicle simulation software ADVISOR, are used to optimize the design of powertrain components. The main outcomes of the research are as follows. The optimization of LCA CO2eq emissions and cost are conflicting as well as cost and energy use, what can be observed in the Pareto solutions. The fuel and LCA CO2eq emissions optimization are coupled for pure hybrids but not for plug-in hybrid configurations, due to the electricity consumption. Fuel cell buses can reduce the energy consumption by 58%, and emit 67% less LCA CO2eq than the conventional diesel bus, and achieve compensatory payback of 0.620 $/km (depending on the hydrogen price). The FC-PHEV configuration shows more potential for achieving higher operation efficiencies, but the FC-HEV shows to have lower life cycle impact and lower cost in general.

This paper presents an overview of the Control, Data Acquisition, and Communication system (CODAC) at the COMPASS tokamak: the hardware set-up, software implementation, and communication tools are described. The diagnostics and the data acquisition are tailored for high spatial and temporal resolution required by the COMPASS physics programme, which aims namely at studies of the plasma edge, pedestal, and Scrape-off-Layer (SOL). Studies of instabilities and turbulence are also an integral part of the programme. Therefore, the data acquisition consists of more than 1000 channels, sampled at rates from 500 kS/s up to 2 GS/s. Presently, the feedback system controls the plasma position and shape, plasma current, and density and it includes 32 analogue input channels as well as 1 digital input/output channel and 8 analogue outputs. The feedback control runs within the Multi-threaded Application Real-Time executor (MARTe) framework with two threads, a 500 μs cycle to control slow systems and a 50 μs cycle to control the fast feedback power supplies for plasma position control. In this paper, special attention is paid to the links between the systems, to the hardware and software connections, and to the communication. The hardware part is described, the software framework is addressed, and the particular implementation – the dedicated software modules, communication protocols, and links to the database are described.

The Advanced Telecommunications Computing Architecture (ATCA) instrumentation platform for fusion control and diagnostics developed at IST, is being extended with a Micro Telecommunications Computing Architecture (MTCA) control and data acquisition platform, still able to maintain performance characteristics in speed, channel density and availability of ATCA, at a cost suitable for smaller systems. Software and firmware have been ported to MTCA using mostly in-house developed hardware, preserving many of ATCA's benefits in performance in a smaller form-factor. The first instance of the proposed platform comprises a MTCA 1U shelf with 6 Advanced Mezzanine Card (AMC) slots, MTCA Carrier Hub (MCH) module, AMC to Field-programmable Gate Array (FPGA) Mezzanine Card (FMC) carrier modules, 500 MSPS to 2 GSPS ADC FMC modules, timing and synchronization input/output (IO) module, or other Commercial Off-The-Shelf (COTS) modules. Data is handled by the MCH, featuring PCIe Gen 3 switching to a host computer, as well as performing clock distribution and Intelligent Platform Management Interface (IPMI) based hardware management for high availability. This compact, flexible and autonomous or scalable platform may either integrate a large-experiment system infrastructure, as well as being configured as a lower-cost, self-contained platform for smaller experiments or development and testbench purposes. The focus of this paper is the development of the MCH MTCA.0 module, which is key to introduce the platform architecture.

Background: The multiplicity of information sources for data acquisition in modern intensive care units (ICUs) makes the resulting databases particularly susceptible to missing data. Missing data can significantly affect the performance of predictive risk modeling, an important technique for developing medical guidelines. The two most commonly used strategies for managing missing data are to impute or delete values, and the former can cause bias, while the later can cause both bias and loss of statistical power. Objectives: In this paper we present a new approach for managing missing data in ICU databases in order to improve overall modeling performance. Methods: We use a statistical classifier followed by fuzzy modeling to more accurately determine which missing data should be imputed and which should not. We firstly develop a simulation test bed to evaluate performance, and then translate that knowledge using exactly the same database as previously published work by [13]. Results: In this work, test beds resulted in datasets with missing data ranging 10-50%. Using this new approach to missing data we are able to significantly improve modeling performance parameters such as accuracy of classifications by an 11%, sensitivity by 13%, and specificity by 10%, including also area under the receiver-operator curve (AUC) improvement of up to 13%. Conclusions: In this work, we improve modeling performance in a simulated test bed, and then confirm improved performance replicating previously published work by using the proposed approach for missing data classification. We offer this new method to other researchers who wish to improve predictive risk modeling performance in the ICU through advanced missing data management.

Patient readmissions to intensive care units (ICUs) are associated with increased mortality, morbidity and costs. Current models for predicting ICU readmissions have moderate predictive value, and can utilize up to twelve variables that may be assessed at various points of the ICU inpatient stay. We postulate that greater predictive value can be achieved with fewer physiological variables, some of which can be assessed in the 24h before discharge. A data mining approach combining fuzzy modeling with tree search feature selection was applied to a large retrospectively collected ICU database (MIMIC II), representing data from four different ICUs at Beth Israel Deaconess Medical Center, Boston. The goal was to predict ICU readmission between 24 and 72h after ICU discharge. Fuzzy modeling combined with sequential forward selection was able to predict readmissions with an area under the receiver-operating curve (AUC) of 0.72+/-0.04, a sensitivity of 0.68+/-0.02 and a specificity of 0.73+/-0.03. Variables selected as having the highest predictive power include mean heart rate, mean temperature, mean platelets, mean non-invasive arterial blood pressure (mean), mean spO2, and mean lactic acid, during the last 24h before discharge. Collection of the six predictive variables selected is not complex in modern ICUs, and their assessment may help support the development of clinical management plans that potentially mitigate the risk of readmission.

This paper proposes a system for the prediction of events in the smart grid. The system infers a label indicating if an event is going to occur in a future time window, in a specific asset, from data of events generated by grid assets and exogenous variables (e.g. weather data). The system design presented follows a sliding-window classification approach, bag-of-words event representation and makes use of random forests models. The systems performance is evaluated in an experimental case study, backed by real data, with the aim of predicting future interruptions in distribution transformers. Performance results indicate that the system is able to deal with highly imbalanced data and validate its adequacy in dealing with the approached problem, achieving up to 0.75 area under the receiver operating characteristic curve in testing.

Patients are often discharged prematurely from Intensive Care Units (ICU) due to clinical resource limitations, economic pressure or poor discharge planning. The readmission of such patients is associated with an increased risk of death and is currently viewed as a marker for poor quality care. Several studies have focused on predicting which patients are likely to be readmitted, using techniques such as logistic regression or machine learning algorithms, and based on physiological data measured during the patients' stay at the ICU. So far, no published algorithms have been able to predict readmissions to a satisfactory degree. In this work we hypothesize that physicians' and nurses' notes could give a better explanation of both ICU discharges and readmissions, and propose using the text notes in an ICU database in order to build classification models for the prediction of readmissions. We tested the use of Fuzzy Fingerprints and other traditional text classifiers and compared them to a previously proposed model based on numerical data, obtaining very relevant improvements in the classification results, namely an AUC=0.8.

This paper proposes a process for the classification of new residential electricity customers. The current state of the art is extended by using a combination of smart metering and survey data and by using model-based feature selection for the classification task. Firstly, the normalized representative consumption profiles of the population are derived through the clustering of data from households. Secondly, new customers are classified using survey data and a limited amount of smart metering data. Thirdly, regression analysis and model-based feature selection results explain the importance of the variables and which are the drivers of different consumption profiles, enabling the extraction of appropriate models. The results of a case study show that the use of survey data significantly increases accuracy of the classification task (up to 20%). Considering four consumption groups, more than half of the customers are correctly classified with only one week of metering data, with more weeks the accuracy is significantly improved. The use of model-based feature selection resulted in the use of a significantly lower number of features allowing an easy interpretation of the derived models.

Probabilistic fuzzy systems (PFS) are shown to be valuable methods for conditional density estimation that combine fuzziness or linguistic uncertainty and probabilistic uncertainty. Several PFS applications have shown the added value of the different reasoning mechanisms of PFS and gains from incorporating two types of uncertainty. The effects of parametrization and parameter estimation on the function or conditional density approximations of PFS have not been documented in the literature. This paper aims to fill this gap in the literature by analyzing the parameters of PFS in conditional density estimation and point forecast using synthetic and real data applications. We show that both in-sample and out-of-sample results depend on PFS parametrization and the results deteriorate when the probability parameters of PFS are not optimized correctly, since these parameters allow the system to be fine tuned.

This paper presents a methodology for short-term load forecasting based on genetic algorithm feature selection and artificial neural network modeling. A feedforward artificial neural network is used to model the 24-h ahead load based on past consumption, weather and stock index data. A genetic algorithm is used in order to find the best subset of variables for modeling. Three datasets of different geographical locations, encompassing areas of different dimensions with distinct load profiles are used in order to evaluate the methodology. The developed approach was found to generate models achieving a minimum mean average percentage error under 2 %. The feature selection algorithm was able to significantly reduce the number of used features and increase the accuracy of the models.

The electricity market has been significantly changing in the last decade. The deployment of smart meters is enabling the logging of huge amounts of data relating to the operations of utilities with the potential of being translated into knowledge on consumers and enable personalized energy efficiency programs. This paper proposes an approach for mining characteristics of a residential consumers (income, education and having children) from high-resolution smart meter data using transparent fuzzy models. The system consists in: (1) extraction of comprehensive consumption features from smart meter data, (2) use of fuzzy models in order to estimate the characteristics of consumers, and (3) knowledge extraction from the fuzzy models rules. Accurate estimates of consumer income and education level were not achieved (60 % accuracy), for the presence of children accuracies of over 70 % were achieved. Performance is comparable to the state of the art with the addition of model interpretability and transparency.

The sampling rate of variables collected in the hospital setting is dependent on several factors. Patients have different lengths of stay in the hospital, during which distinct physiological parameters are measured. The frequency of measurements depends ultimately in the type of variable and in the patient condition. Hence, when performing data based modeling for knowledge discovery in medical databases, one should have in consideration the heterogeneity of variables. This paper proposes an extension of a mixed fuzzy clustering algorithm in order to handle time invariant and time variant features of unequal lengths. Additionally, a novel approach for deriving Takagi-Sugeno fuzzy models, based on feature transformation using fuzzy c-means is implemented and compared with approaches based on mixed fuzzy clustering. The proposed approaches are tested on real data for mortality prediction in intensive care units of patients diagnosed with acute kidney injury and for ICU readmission prediction. Overall, mixed fuzzy clustering yields better results than fuzzy c-means. Moreover, the proposed extension for time series of unequal lengths improves previous results. Mortality is classified with an AUC of 0.73 and readmissions with an AUC of 0.64.

In many binary medical classification problems, the cost of misclassifying one category is higher than the other, and in these applications it is desirable to employ a classifier with selective sensitivity or specificity. This work explores the utility of a fuzzy multi-criteria function for performance evaluation during knowledge-based medical classification and prediction. The method presented here uses fuzzy optimization to combine the sensitivity, specificity, and accuracy of classification as goals in a single objective function. This approach is used to assign flexible goals, which can be used to maximize the outcome in terms of each one of the goals. The proposed approach significantly increases the sensitivity and the specificity while maintaining or increasing accuracy. The versatility of the method is further exploited in a multi-model approach, using individual structures of multi-objective optimization of sensitivity and specificity separately, and then combining their outcomes through a decision-making module. Among various medical benefits derived from applying this technique, the divergent feature sets selected by high sensitivity and specificity models lend insight into factors more integrally connected to what causes risk of death for patients.

In this paper, the prime importance of preprocessing in clinical databases is discussed. Specifically in intensive care units, data is often irregularly recorded, contain a large amount of missing values and sampling times are uneven. This paper proposes a systematic preprocessing procedure that can be generalized to common clinical databases. This procedure is applied to a known septic shock patient database and classification results are compared with previous studies. The goal is to estimate, as accurately as possible, the outcome (survived or deceased) of these septic shock patients. Neural modeling is used for classification. Detailed classification results are presented and show that the preprocessing is crucial to improve classifiers accuracy.

This work introduces a fuzzy multi-objective function for performance evaluation in feature selection for a knowledge based medical classification task. Fuzzy optimization was used to combine the sensitivity, specificity, and percentage of correct classifications as objectives, and assign flexible goals to strengthen the sensitivity. Using the same soft computing methods and publicly available septic shock database as in Fialho et al. (2010), an increase in sensitivity of 30% was achieved while maintaining the percentage of correct classifications. Feature subsets selected were different from those previously published, lending insight into factors more integrally connected to what causes risk of death for patients.

This paper proposes the application of a new wrapper particle swarm optimization (PSO) method to feature selection problems. The new method was designed to cope with premature convergence for the binary PSO algorithm. Its principal objective is to improve classification accuracy of fuzzy models. Further, the method is tested in a medical database, namely in the prediction of the outcome (survived or deceased) of sepsis patients.

The available set of potential features in real-world databases is sometimes very large, and it can be necessary to find a small subset for classification purposes. One of the most important techniques in data pre-processing for classification is feature selection. Less relevant or highly correlated features decrease, in general, the classification accuracy and enlarge the complexity of the classifier. The goal is to find a reduced set of features that reveals the best classification accuracy for a classifier. Rule-based fuzzy models can be acquired from numerical data, and be used as classifiers. As rule based structures revealed to be a useful qualitative description for classification systems, this work uses fuzzy models as classifiers. This paper proposes an algorithm for feature selection based on two cooperative ant colonies, which minimizes two objectives: the number of features and the classification error. Two pheromone matrices and two different heuristics are used for these objectives. The performance of the method is compared with other features selection methods, achieving equal or better performance.

Bayesian networks (BNs) are knowledge representation tools capable of representing dependence or independence relationships among random variables. Learning the structure of BNs from datasets has received increasing attention in the last two decades, due to the BNs' capacity of providing good inference models and discovering the structure of complex domains. One approach for BNs' structure learning from data is to define a scoring metric that evaluates the quality of the candidate networks, given a dataset, and then apply an optimization procedure to explore the set of candidate networks. Among the most frequently used optimization methods for BN score-based learning is greedy hill climbing (GHC) search. This paper proposes a new local discovery ant colony optimization (ACO) algorithm and a hybrid algorithm max-min ant colony optimization (MMACO), based on the local discovery algorithm max-min parents and children (MMPC) and ACO to learn the structure of a BN. In MMACO, MMPC is used to construct the skeleton of the BN and ACO is used to orientate the skeleton edges, thus returning the final structure. The algorithms are applied to several sets of benchmark networks and are shown to outperform the GHC and simulated annealing algorithms.

Model-based fault detection and isolation (FDI) is an approach with increasing attention in the academic and industrial fields, due to economical and safety related matters. In FDI, the discrepancies between system outputs and model outputs are called residuals, and are used to detect and isolate faults. This paper proposes a model-based architecture for fault detection and isolation based on fuzzy methods. Fuzzy modeling is used to derive nonlinear models for the process running in normal operation and for each fault. When a fault occurs, fault detection is performed using the residuals. Then, the faulty fuzzy models are used to isolate a fault. The FDI architecture proposed in this paper uses a fuzzy decision making approach to isolate faults, which is based on the analysis of the residuals. Fuzzy decision factors are derived to isolate faults. An industrial valve simulator is used to obtain several abrupt and incipient faults, which are some of the possible faults in the real system. The proposed fuzzy FDI architecture was able to detect and isolate the simulated abrupt and incipient faults.

This paper proposes input selection methods for fuzzy modeling, which are based on decision tree search approaches. The branching decision at each node of the tree is made based on the accuracy of the model available at the node. We propose two different approaches of decision tree search algorithms: bottom-up and top-down and four different measures for selecting the most appropriate set of inputs at every branching node (or decision node). Both decision tree approaches are tested using real-world application examples. These methods are applied to fuzzy modeling of two different classification problems and to fuzzy modeling of two dynamic processes. The models accuracy of the four different examples are compared in terms of several performance measures. Moreover, the advantages and drawbacks of using bottom-up or top-down approaches are discussed.

This papers proposes two novel approaches for the identification of Takagi-Sugeno fuzzy models with time variant and invariant features. The proposed Mixed Fuzzy Clustering algorithm is proposed for determining the parameters of Takagi-Sugeno fuzzy models in two different ways: (1) the antecedent fuzzy sets are determined based on the partition matrix generated by the Mixed Fuzzy Clustering algorithm; (2) the input features are transformed using the same algorithm and the antecedent fuzzy sets are derived using Fuzzy C-Means clustering. The proposed approaches are tested on four different health care applications: readmissions in intensive care units, administration of vasopressors and mortality. The results show that the proposed clustering algorithm resulted in an increase of the performance of the fuzzy models in three out of four applications in comparison to the use of Fuzzy C-Means.

The current demands on the energy market, such as efficiency, sustainability and affordability increase the need for customer understanding and data analysis. This paper presents an analysis of the segmentation of electricity consumers based on the fuzzy clustering of time variant electricity consumption data and invariant features like the demographic customer information. The algorithm used is mixed fuzzy clustering (MFC), which allows integrating both variant and invariant features into one clustering. The clustering is evaluated both in its stability over the two years of data, using a entropy measurement and in its general quality given by the three clustering validity indices, Calinski-Harabasz, Davies-Bouldin and Silhouette index.

This work proposes a methodology for predicting the typical daily load profile of electricity usage based on static data obtained from surveys. The methodology intends to: (1) determine consumer segments based on the metering data using the k-means clustering algorithm, (2) correlate survey data to the segments, and (3) develop statistical and machine learning classification models to predict the demand profile of the consumers. The developed classification models contribute to make the study and planning of demand side management programs easier, provide means for studying the impact of alternative tariff setting methods and generate useful knowledge for policy makers.

The electricity market has been significantly changing in the last decade. The deployment of smart meters is enabling the logging of huge amounts of data relating to the operations of utilities with the potential of being translated into valuable knowledge on the behaviour of consumers. This work proposes a methodology for predicting the typical daily load profile of electricity usage based on static data using fuzzy clustering and modelling. The methodology intends to: (1) determine consumer segments based on the metering data using the fuzzy c-means clustering algorithm, and (2) develop Takagi-Sugeno fuzzy models in order to predict the demand profile of the consumers.

Several algorithms have been proposed in the past to solve the problem of binary pattern recognition. The problem of finding features that clearly distinguish two or more different patterns is a key issue in the design of such algorithms. In this paper, a graph-like recognition process is proposed that combines a number of different classifiers to simplify the type of features and classifiers used in each classification step. The graph-like classification method is applied to ancient music optical recogniti on, and a high degree of accuracy has been achieved.

The aim of this work is to identify groups of patients with similar patterns that are related to a higher risk of readmission to an Intensive Care Unit (ICU). Patients readmissions to ICUs are introduced as a problem associated with increased mortality, morbidity and costs, which complicates the performance of a good clinical management and medical diagnosis. To approach the readmissions classification problem, Fuzzy C-Means (FCM) clustering algorithm was implemented to find the different groups of patients. A multimodel approach was developed using these groups and the best clustering division was assessed through different objective functions. Two decision criteria were used for the multimodel approach, an a priori decision and an a posteriori decision. The data used, from MIMIC II database, consisted on the arithmetic means of time series of variables - acquired during the last 24 hours before discharge. The multimodel using the a priori and the aposteriori decisions were able to predict readmissions with an average AUC of 0.74 and 0.75, respectively. Consequently, the multimodel results overcame the results of previous predictive models developed for the classification of readmissions outcome.

This paper proposes a novel feature selection approach formulated based on the Fish School Search (FSS) optimization algorithm, intended to cope with premature convergence. In order to use this population based optimization algorithm in feature selection problems, we propose the use of a binary encoding scheme for the internal mechanisms of the fish school search, emerging the binary fish school search (BFSS). The suggested algorithm was combined with fuzzy modeling in a wrapper approach for Feature Selection (FS) and tested over three benchmark databases. This hybrid proposal was applied to an ICU (Intensive Care Unit) readmission problem. The purpose of this application was to predict the readmission of ICU patients within 24 to 72 hours after being discharged. We assessed the experimental results in terms of performance measures and the number of features selected by each used FS algorithms. We observed that our proposal can correctly select the discriminating input features.

Shock is a life-threatening medical condition requiring the administration of powerful drugs - vasopressors. Early identification of these patients is a worthy goal in order to timely prepare them for therapy. A subset composed of the most frequently sampled and readily available variables in an intensive care unit (ICU) was used for clustering patients. Then, a data exploration process was started through the use of fuzzy clustering with the fuzzy cmeans algorithm, where four clusters were obtained and the groups characteristics were analyzed. A relationship between the clusters obtained and the use of vasopressors was found out and these results were visualized with the help of histograms. First, a single model was derived. Then, four models were trained and used for a multi model approach, one for each identified group of patients. In both cases fuzzy models were used as they are universal approximators. For the multi-model approach, two decision criteria were used. First a decision a priori based on the distance from the clusters centers to the patient characteristics was used. Lastly a decision a posteriori approach where each model was used and the final outcome used is based on the uncertainty of the output response to the threshold of each model. The multi model approach with a posteriori decision had a better performance of the two schemes tested, and also performed better than the single general model approach.

We propose the application of probabilistic fuzzy systems (PFS) to model the prediction of early readmission in intensive care unit patients and compare it with the gold-standard method - logistic regression based on the APACHE II score. PFS are characterized by the combination of the linguistic description of the system with the statistical properties of data. On one hand, results point that PFS models perform comparably to the gold-standard method, with AUC values of 0.66±0.03. On the other hand, results also show that PFS models use a significant lower number of variables which, from the clinical practice point of view, suggests improved gains in terms of simplicity.

Medical databases often contain large amounts of missing data. This poses very strict constraints to the use of exclusively computer-based feature selection techniques. Moreover, in medical data there is usually no unique combination of features that provides the best explanation of the outcome. In this paper we propose a modified Sequential Forward Selection (SFS) approach to the problem of selecting sets of physiologic variables from septic shock patients in order to predict their outcome. We were able to achieve ten different combinations of only three physiological numerical parameters, all performing better than the best set suggested up to now. The performances of these sets are higher than 0.97 for AUC and up to 0.97 for accuracy.

Soft computing, as an engineering science, and statistics, as a classical branch of mathematics, emphasize different aspects of data analysis. Soft computing focuses on obtaining working solutions quickly, accepting approximations and unconventional approaches. Its strength lies in its flexibility to create models that suit the needs arising in applications. In addition, it emphasizes the need for intuitive and interpretable models, which are tolerant to imprecision and uncertainty. Statistics is more rigorous and focuses on establishing objective conclusions based on experimental data by analyzing the possible situations and their (relative) likelihood. It emphasizes the need for mathematical methods and tools to assess solutions and guarantee performance. Combining the two fields enhances the robustness and generalizability of data analysis methods, while preserving the flexibility to solve real-world problems efficiently and intuitively.

The amount of data generated in the intensive care environment nowadays prohibits the storage of all the information available. The validation process is time consuming, since nurses have to check every certain periods the data acquired from bedside monitors in order to assess their validity and integrity. This work presents an automatic method for data validation in the intensive care environment, based on an artificial intelligence approach, namely artificial neural networks (ANNs). A real world dataset acquired at Beth Israel Deaconess Medical Center (BIDMC) neonatal intensive care unit (NICU) is used to obtain the validation model and assess its performance. The dataset consists of high frequency sampled data of the level of oxygen saturation (SpO2) of neonates. A subset of 100 neonates was considered for modeling purposes. A total of 7,018,662 samples were available, containing 129,075 validated ones. The performance of the validation model, assessed in terms of its AUC, was of up to 0.75. Both the sensitivity and specificity reached acceptable values according to medical review. Future work would involve a prospective study and validation of the methods proposed in this work.

This paper proposes the application of a new binary particle swarm optimization (BPSO) method to feature selection problems. Two enhanced versions of binary particle swarm optimization, designed to cope with premature convergence of the BPSO algorithm, are proposed. These methods control the swarm variability using the velocity and the similarity between best swarm solutions. The proposed PSO methods use neural networks, fuzzy models and support vector machines in a wrapper approach, and are tested in a benchmark database. It was shown that the proposed BPSO approaches require an inferior simulation time, less selected features and increase accuracy. The best BPSO is then compared with genetic algorithms (GA) and applied to a real medical application, a sepsis patient database. The objective is to predict the outcome (survived or deceased) of the sepsis patients. It was shown that the proposed BPSO approaches are similar in terms of model accuracy when compared to GA, while requiring an inferior simulation time and less selected features.

Real-world databases often contain missing data and existing correction algorithms deliver varying performance. Also, most modeling techniques are not suitable to deal with them automatically. In this study we examine different approaches to predicting septic shock in the presence of missing data. Some preprocessing techniques for managing missing data include disregarding data, or replacing it with information that by design introduces bias. In this study, we show that predictive performance improves by employing a minimum pre-processing technique, the Zero-Order-Hold (ZOH) method, by applying a Fuzzy C-Means clustering technique based on the partial distance calculation strategy (FCM-PDS) and by computing the final classification regarding the samples from each patient. Performance improvements continue to occur where up to approximately 60% of the data is missing, though for higher percentage the classification performance still is statistically improved. We further validate this approach by making comparisons with previous studies.

Laboratory testing is a frequent activity for patients in intensive care units (ICU). Recent studies demonstrate that frequent laboratory testing does not necessarily relate to better outcomes. We hypothesize that unnecessary laboratory testing can be reduced by predicting which tests are unlikely to influence clinical management. Reducing unnecessary tests could reduce morbidity and hospitalization costs. We analyzed an ICU database containing 26,665 patient records at Beth Israel Deaconess Medical Center, Boston, and selected a subset of patients with gastrointestinal bleeding. Database knowledge discovery was applied involving data preprocessing, feature selection, and classification. Conventional soft computing tools such as fuzzy models and neural networks were utilized in this work, combined with statistical and mathematical tools. The input variables included bedside monitor trends, lab tests, arterial/central catheter information, urine collections, transfusions, indexes and scores calculated for the patients. The outcome variable was a binary classification based on falling levels of hematocrit. Feature selection was performed by a bottom-up strategy, maximizing the area under the ROC curve (AUC), the integrated discrimination improvement (IDI) and a multiobjective function primarily pondering the sensitivity of the models. A leave-one-out cross validation process was used to evaluate the overall models' performance, as well as the additional predictive value of the variables selected. Urine output was selected by all models as the best predictor of useful hematocrit testing. Our results show that it is possible to correctly classify the usefulness of a hematocrit lab test up to 81% of the time by using fuzzy models and neural networks.

SAFE-PORT is a recently started project addressing the complex issue of determining the best configurations of resources for harbour and port surveillance and protection. More specifically, the main goal is to find, for any given scenario, an adequate set of configuration solutions - i.e., number and type of sensors and equipments, their locations and operating modes, the corresponding personnel and other support resources - that maximize protection over a specific area. The project includes research and development of sensors models, novel algorithms for optimization and decision support, and a computer-based decision support system (DSS) to assist decision makers in that task. It includes also the development of a simulation environment for modelling relevant aspects of the scenario (including sensors used for surveillance, platforms, threats and the environment), capable to incorporate data from field-trials, used to test and validate solutions proposed by the DSS. Test cases will consider the use of intelligent agents to model the behaviour of threats and of NATO forces in a realistic way, following experts' definitions and parameters.

This paper presents a slow control system based on Experimental Physics and Industrial Control System (EPICS) and FireSignal technology for fusion devices, that is being developed for ISTTOK and COMPASS tokamaks. In this approach, the machine sub-systems control is based in micro-controllers. These sub-systems can handle automatically the necessary procedures before, during and after a plasma discharge, including safety protocols. An EPICS based supervisory control software then integrates all machine subsystems for central monitoring, configuration, logging and alarm management. Instead of the usual EPICS MEDM client tools, remote monitoring and configuration is accomplished by means of a specifically developed Java application using the CAJ library for EPICS and Common Object Request Broker Architecture (CORBA). This application behaves also as FireSignal Client, allowing the subsystems to be completely integrated on a FireSignal-based CODAC with all Process Variables (PV) and relevant control events being recorded in the central database. Through the FireSignal operator console, the so called FS User Client, it is possible to present a dedicated GUI for each subsystem, thus providing the machine operator a complete overview of the state of the machine and full control capabilities.

Real word data sets often contain many missing elements. Most algorithms that automatically develop a rule-based model are not well suited to deal with incomplete data. The usual technique is to disregard the missing values or substitute them by a best guess estimate, which can bias the results. In this paper we propose a new method for estimating the parameters of a Takagi-Sugeno fuzzy model in the presence of incomplete data. We also propose an inference mechanism that can deal with the incomplete data. The presented method has the added advantage that it does not require imputation or iterative guess-estimate of the missing values. This methodology is applied to fuzzy modeling of a classification and regression problem. The performance of the obtained models are comparable with the results obtained when using a complete data set.

This paper proposes the application of new knowledge based methods to a septic shock patient database. It uses wrapper methods (bottom-up tree search or ant feature selection) to reduce the number of features. Fuzzy and neural modeling are used for classification. The goal is to estimate, as accurately as possible, the outcome (survived or deceased) of these septic shock patients. Results show that the approaches presented outperform any previous solutions, specifically in terms of sensitivity.

Measuring the performance of a given classifier is not a straightforward or easy task. Depending on the application, the overall classification rate may not be sufficient if one, or more, of the classes fail in prediction. This problem is also reflected in the feature selection process, especially when a wrapper method is used. Cohen's kappa coefficient is a statistical measure of inter-rater agreement for qualitative items. It is generally thought to be a more robust measure than simple percent agreement calculation, since it takes into account the agreement occurring by chance. Considering that kappa is a more conservative measure, then its use in wrapper feature selection is suitable to test the performance of the models. This paper proposes the use of the kappa measure as an evaluation measure in a feature selection wrapper approach. In the proposed approach, fuzzy models are used to test the feature subsets and fuzzy criteria are used to formulate the feature selection problem. Results show that using the kappa measure leads to more accurate classifiers, and therefore it leads to feature subset solutions with more relevant features.