International Joint Conference SOCO’17-CISIS’17-ICEUTE’17 León, Spain, September 6–8, 2017, Proceeding

In the educational services, students’ emotions are an important factor that determine its effect. We have previously conducted research that led them to target emotions using environmental factors. However, the study used the bayesian network based on domain knowledge to predict emotions, which may differ from the actual environment. In this paper, we propose a method to learn the bayesian network for group emotion prediction in kindergarten from data through evolutionary computation. The learning data are brightness, color temperature, sound, volume, smell, temperature, humidity, and current emotion. The structure of the network is encoded with two chromosomes to represent nodes and arcs. To explore the optimal structure, evolutionary operators are used that can convey information in sets. We also experiment with various inference nodes not observed. Experimental results show that the accuracy is 85% with 20 inference nodes, which can replace network designed with domain knowledge. By comparing the evolution of the best model, we analyze the influential factors that determine the structure.

Modern composite scientific applications, also called scientific workflows, require large processing capacities. Cloud environments provide high performance and flexible infrastructure, which can be easily employed for workflows execution. Since cloud resources are paid in the most cases, there is a need to utilize these resources with maximal efficiency. In this paper we propose dynamic resources coevolutionary genetic algorithm, which extends previously developed coevolutionary genetic algorithm for dynamic cloud environment by changing computational capacities of execution nodes on runtime. This method along with using two types of chromosomes – mapping of tasks on resources and resources configuration – allows to greatly extend the search space of the algorithm. Experimental results demonstrate that developed algorithm is able to generate solutions better than other scheduling algorithms for a variety of scientific workflows.

Berge-Zhukovskii equilibrium is an alternate solution concept to Nash equilibrium that induces cooperation in non-cooperative games. A solution of a game is a Berge-Zhukovskii equilibrium if the payoff of each player cannot increase regardless of what the other players do. The Berge-Zhukovskii equilibrium has been found to be us useful in trust games. We propose a new method, based on evolutionary algorithms, to detect and track the Berge-Zhukovskii equilibrium of a game considering a discrete-time dynamic environment. To test our method we propose a new dynamic multiplayer game model, based on the Voluntary contribution mechanism. Numerical results show the potential of the proposed method.

Zeolites are crystalline materials widely used in many catalytic process in industry. Specifically, they have a major impact at petrochemicals, fine chemicals or gas separation. Thus, discovering new zeolites with specific properties is a high-impact target for the industry, due to their huge economical repercussions. New tools and techniques are needed to help in this task, because trial and error approaches prevail until now. In this work, we propose a new application of genetic algorithms for helping chemical engineers in the process of determining zeolite structures with specific properties. Our proposal takes advantage of some symmetry operation properties to improve the performance of the genetic algorithm. Furthermore, a suitable fitness function has been designed which evaluates all main features required for efficient zeolites.

An empirical comparison of different intelligent soft sensors for obtaining the state of health of automotive rechargeable batteries is presented. Data streamed from on-vehicle sensors of current, voltage and temperature is processed through a selection of model-based observers of the state of health, including data-driven statistical models, first principle-based models, fuzzy observers and recurrent neural networks with different topologies. It is concluded that certain types of recurrent neural networks can outperform well established first-principle models and provide the supervisor with a prompt reading of the State of Health. The algorithms have been validated with automotive Li-FePO$$_4$$ cells.

Intelligent Systems are the best way to manage complex industrial processes with a high number of process parameters, like injection molding process. Specifically, Fuzzy Logic is a solution to estimate if the qualitative inspection of parts produced allows us to determine correct process parameter value setting to produce good quality parts. This paper shows an intelligent decision system based on Fuzzy Logic techniques designed using defect behavior tendency curves as membership functions. These functions are improving with dynamics and adaptive regression membership functions based on the assessment of quality for a given part done by an operator. The implementation of this intelligent decision system designed for injection molding process shows that is able to transform a qualitative variable deduced of qualitative injection inspection of part defects, into a quantitative inspection, identifying the correct process parameters. Experimental results show that the effectiveness is improved and also reduces the time of a process in a 40%.

This work presents a hardware and software solution that implements algorithms based on intelligent computing techniques for estimating the stress level using low cost platforms. These algorithms process the acquired physiological signals directly from the sensors using advanced filtering and processing techniques and algorithms based on fuzzy logic. For this purpose, a hardware configuration based on the Arduino Uno and Raspberry Pi 3 platforms has been chosen. These platforms perform the acquisition, processing and upload of the data to a server via WiFi. In the implementation of the server a configuration based on Linux, Apache, MySQL and PHP (LAMP) has been carried out. The parameters used to estimate the stress level derive from the following physiological signals: the electrocardiogram (ECG) and the galvanic skin response (GSR).

Medical staff performance and staff scheduling represent a significant determinant of public healthcare quality where the cost effectiveness is required from hospitals. There are two cases regarding the nurse scheduling problem, the static and dynamic. Dynamic nurse scheduling problem, often called nurse rerostering problem, which presents reconstruction or modification of the predetermined roster for the current scheduling horizon. This paper is focused on a new strategy of hybrid approach based on fuzzy ordered weighted averaging empirical model for detecting the best solution in nurse rerostering problem. The model is tested with original real-world dataset obtained from the Oncology Institute of Vojvodina in Serbia.

This paper proposes a problem formulation to aid as a support information management system of a wind power producer having energy storage devices and participating in electricity markets. Energy storage can play an important role in the reduction of uncertainties faced by a wind power producer. Excess of conversion of wind energy into electric energy can be stored and then released at favorable hours. Energy storage provides capability for arbitrage and increases the revenue of the wind power producers participating in electricity markets. The formulation models the wind power and the market prices as stochastic processes represented by a set of convenient scenarios. The problem is solved by a powerful stochastic mixed integer linear programming problem. A case study using data from the Iberian Electricity Market is presented to show the aid of the formulation.

The objective of this work is to improve gas supply efficiency in residential districts. To achieve this goal, Artificial Neural Networks (ANNs) have been used. In this work, a hybrid model based on ANN has been proposed that obtains total daily gas consumption (in KWh) in residential districts, with a prediction horizon of 7 days. Previous consumption records and meteorological variables have been considered to improve the prediction of future gas consumption. In order to find the best ANN that models the behavior of this consumption variable, a set of experiments has been designed, where the mean square error of each network is measured to rate their reliability and accuracy. A hybrid neural model has been created to determine a horizon of 7 predictions using a median filter of the 5 best predictors per day.

The Heat Pump with geothermal exchanger is one of the best methods to heat up a building. The heat exchanger is one of the most representative elements when a heat pump is employed as building heating system. In the present study, a novel intelligent system was designed to predict the performance of on this kind of heating equipment. The novel approach has been successfully empirically tested under a real dataset obtained during measurements along one year. It was based on time series modeling. Then, the model was validated and verified; it obtains good results in all the operating conditions ranges.

A key challenge in electrical discharge machining (EDM) is to find a suitable combination out of numerous process parameters. Any changes, concerning the electrode materials or geometries, require newly optimized technologies. These technologies are to be developed from a considerable number of experiments which must be carried out by an experienced operator.This paper presents a new method of finding the optimal parameters.It seems likely that the evolution strategy (ES), a stochastic, metaheuristic optimization method, offers the great advantage of finding solutions, even with little knowledge of system behaviour.The method involved a randomized and a derandomized ES, based on a non-elitistical (μ,λ)-evolution strategy with one parent and four children. The two ES were initialized from an unfavourable starting point (A) and from a favourable starting point (B), to investigate their effectiveness.We demonstrate that starting from the unfavourable starting point A the processing duration t_ero could be reduced by a maximum of 77% with a slightly smaller linear wear of the tool electrode Δl_E after 40 trials.

A new method for detecting cardiac arrhythmias is proposed. The differences between the instantaneous frequencies of signals recorded in atrium and ventricle are computed by means of a non-linear spectral transform. This transform dilates or contracts the time scale until the ventricle signal has a flat frequency spectrum in time. Singular Spectrum Analysis is used to isolate its oscillatory components. The same temporal dilations and contractions are applied to the atrium signal, that is subsequently projected onto the oscillatory components found in the ventricle signal. It is shown that the frequency spectrum of the processed atrium signal becomes uneven only at arrhythmia episodes.

Modern cities are a subject for various threats like terrorist attacks or natural disasters. Effective response on them requires fast delivering of information as close as possible to the source of events. Online social networks can play a role of monitoring system for such kind events with its users as particular sensors. But to exploit such a system one requires to have capabilities to process noisy, distorted data where desired information represented as compound entities scattered across the text of users’ messages, consisting of specific keywords, location names and service words heavily affected by usage styles of online social networks. This paper presents effective approach to handle with the problem of compound location extraction from online social networks in Russian language.

The correct diagnosis of tool breakage is fundamental to improve productivity, minimizing the number of unproductive hours and avoiding expensive repairs. The use of Data Mining techniques provides a significant added value in terms of improvements in the robustness, reliability and flexibility of the monitored systems. In this work, a general view of a diagnosis and prognosis of tool breakage in Industrial Processes is proposed. The important issues identified will be analyzed: filtering, process characterization and data based modeling. A case study has been implemented to carry out the prognosis of tool breakage in the cold stamping process. The results provided are qualitative trends and hypothesis to perform the prognosis. Although a validation in real operation is needed, these results are promising and demonstrate the goodness of using these type of techniques in real processes.

Nowadays, both students performance and its evaluation are important challenges and play a significant role, in general terms. Frequently, the students attempts to pass a specific curriculum subjects, have several fails due to different reasons and, in this context, lack of data adversely affects interesting future analysis for achieving conclusions. As a consequence, data imputation processes must be performed in order to substitute the missing data for estimated values. This paper presents a comparison between two data imputation methods developed by the authors in previous researches, the Adaptive Assignation Algorithm (AAA) based on Multivariate Adaptive Regression Splines (MARS), and the Multivariate Imputation by Chained Equations methodology (MICE). The results obtained demonstrate that both proposed methods achieve good results, specially AAA algorithm.

This study is focused on achieving a reliable prediction of the daily number of goods subject to inspection at Border Inspections Posts (BIPs). The final aim is to develop a prediction tool in order to aid the decision-making in the inspection process. The best artificial neural network (ANN) model was obtained by applying the Bayesian regularization approach. Furthermore, this study compares daily forecasting with a two-stage forecasting approach using a weekly aggregation-disaggregation procedure. The comparison was made using different performance indices. The BIP of the Port of Algeciras Bay was used as a case study. This approach may become a supporting tool for the prediction of the number of goods subject to inspection at other international inspection facilities.

The systems monitoring the location of public transport vehicles rely on wireless transmission. The location readings from GPS-based devices are received with some latency caused by periodical data transmission and temporal problems preventing data transmission. This negatively affects identification of delayed vehicles. The primary objective of the work is to propose short term hybrid delay prediction method. The method relies on adaptive selection of Hoeffding trees, being stream classification technique and multilayer perceptrons. In this way, the hybrid method proposed in this study provides anytime predictions and eliminates the need to collect extensive training data before any predictions can be made. Moreover, the use of neural networks increases the accuracy of the predictions compared with the use of Hoeffding trees only.

Construction and use of software agents for industrial and business computer systems is a well-known subject for professional development teams. But the full potential of Agent-Oriented programming usually remains hidden to these groups and agents are usually not exploited to their full potential. This study shows the description and implementation of a number of techniques created to face the construction of agents composed of a number of behaviors that may be debugged and revised in a step-by-step fashion, both at design and at runtime. The set of techniques also include a mechanism for managing the orchestration of behaviors, what gives the development teams insight on the final behavior resulting from the combination of the individual ones. These techniques allow the application of agile software development principles to multiagent environments, making it easier to integrate these kind of software artifacts into complex projects using this paradigm. The paper also describes two practical sample use cases where these techniques are employed: the application of the techniques to the simple JADE DummyAgent and to complex rule-based agents that run behaviors based on CLIPS or JESS rules.

The concept of community in complex networks, which is intuitively expressed as a group of nodes more densely connected to each other than to the rest of the network, has not been formally defined yet in a manner that encompasses all aspect ensuing from this intuitive description. Among existing approaches, a popular one consists in considering the network community structure as the optimum value of a fitness function that reflects the modularity of the network. Recently, a new trend to model the problem as a game having the community structure as equilibrium has emerged. Both approaches are appealing as they allow the design of heuristic approaches to this problem and benefit from their adaptability and scalability. This paper analyzes the behavior of such a heuristic that is based on extremal optimization in combination with two possible game theoretic models that consider different payoff functions, in comparison with the corresponding optimization approaches.

This paper proposes an innovative single point cutting device that requires less maintenance than traditional micro-milling machines, being the cutting tools required simpler easier to develop. This satisfies the market demands on micro manufacturing, where devices must be more accurate and cheaper, able to create a diverse range of shapes and geometries with a high degree of accuracy. A stereo vision system has been implemented as an alternative to the current technologies in the market to locate the tool tip, and with this, make the proper corrections on the CNC machine. In this paper the development of such system is explored and discussed. Experimental results show the accuracy of the proposed system, given an error in the measurement of $${\pm }3\,{\upmu }\mathrm{m}$$.

This contribution introduces ICARO-3D, an on going research project which is focused on developing computer-based methods for the automatic reconstruction and categorization/classification of archaeological remains. The system provides a pipeline facing the two main stages of the procedure, a.k.a the 3D reconstruction and the categorization of archaeological remains. In this work, we describe the approach adopted in the latter stage. While our 3D reconstruction procedure provides a 3D model using Soft computing techniques, the categorization of the modeled remains deals with exploiting the descriptive power of the well-known Viewpoint Feature Histogram (VFH) descriptor. The conducted experiments reveal us the promising results of the considered approach. Moreover, our contribution follows a reproducible research (RR) policy in order to guarantee the integrity of the reported results. Then, both the source code and the image datasets are accordingly available.

Patents can become indicators of various social, economic or technological aspects, and this is why an analysis of these documents is of our interest. This paper proposes a platform that retrieves, analyses and visualizes functionalities that represent data on the landscape of patents obtained from the Spanish Patent and Trademark Office (OEPM).

Two different models for instability detection in a radial turning process are proposed in order to prevent fault appearance. This methods allows to detect instability on this machining process based on the forces. Median Absolute Deviation Normalized (MADN) and Principal Component Analysis (PCA) are the statistical methods used to classify those tests. The results have showed that the models are close to expert classification of the tests stability.

This paper reports the use of the PAELLA algorithm in the context of weighted regression. First, an experiment comparing this new approach versus probabilistic macro sampling is reported, as a natural extension of previous work. Then another different experiment is reported where this approach is tested against a state of the art regression technique. Both experiments provide satisfactory results.

With the advances in Language Technologies and Natural Language Processing, conversational interfaces have begun to play an increasingly important role in the design of human-machine interaction systems in a number of devices and intelligent environments. In this paper, we present a statistical model for spoken dialog segmentation and labeling based on a generative model learned using decision trees. We have applied our proposal in a practical conversational system that helps solving simple and routine software and hardware repairing problems. The results of the evaluation show that automatic segmentation of spoken dialogs is very effective for human-machine dialogs. The same statistical model has been applied to human-human conversations and provides a good baseline as well insights in the model limitation.

The possibility to simulate hypothetical scenarios without harming the business using What-If analysis tools and to retrieve highly refined information is an interesting way of achieving such advantages. In a previous work, it was introduced a hybrid model that combines What-If analysis and OLAP usage preferences, which helps filtering the information, meeting the users’ needs and business requirements without losing data quality. In addition, it helps to overcome the lack of a user expertise using What-If analysis process. In this paper, we propose a formal verification of a hybridization model, integrating What-If analysis scenarios with OLAP usage preferences, which aim to suggest to the user enriched What-If scenarios based on the usage preferences of a specific user. For this, we used Alloy to specify and verify the referred model, which enables to analyze and verify our hybrid model, discovering possible ambiguity and inconsistencies.

This study presents the application of dimensionality reduction and clustering techniques to episodes of high pollution in Madrid City (Spain). The goal of this work is to compare two scenarios with similar atmospheric conditions (periods of high NO₂ concentration): one of them when no actions were taken and the other one when traffic restrictions were imposed. The analyzed data have been gathered from two acquisition stations from the local air control network of Madrid City. The main pollutants recorded at these stations along four days during two time intervals are analyzed in order to determine the effectiveness of the anti-pollution measures. Dimensionality-reduction and clustering techniques have been applied to analyse the pollution public datasets.

After concisely reviewing the related work, the job shop transient scheduling problem and the model of production system of repetitive production are discussed. The activity-oriented Petri net (AOPN) simulation approach for optimization of transient dispatching rules is given. The software tool GPenSim and computational experiments with the results are discussed. The simulation results show that by the AOPN approach and with the GPenSIM implementation, it is possible to run experiments with various transient logic. The paper concludes with a summary and directions for future work.

This paper presents the analysis of an approach to the problem Virtual Manufacturing (VM) networks coordinating utilizing some concepts from the area of artificial intelligence. VM network consists of group of manufacturers and their resources that are flexible utilized to fulfill the common manufacturing task. In the proposed paper it is analyzed the problem of coordinating such VM network using the multi-agent system (MAS) concept and the multiple ant colony (MAC) approach. In the presented concept the agents are defined as virtual ants in the MAC approach. Results obtained during simulation tests let to determine main advantages of this approach.

It becomes increasingly complex to make decisions concerning environmental requirements at the product design stage. Any decision made at this stage may affect the environmental impact of the product. It is therefore necessary to provide new tools for product designers to help them in ecological design choices. This paper describes a new method for ecodesigning technological processes. The method was implemented in an expert system, and it supports process engineers in the development of eco-friendly technological processes. Decision tree induction method was used. The environmental approach to designing technological processes described in this paper is a new solution, and it automates the selection of materials and connections in the designed product.

Communicating agents are an important part of intelligent machines. However, modeling communicating agent with Petri nets is not an easy task as the modeling usually yields large yet less expressive models. In this paper, a modular Petri net approach is introduced for modeling communicating agents. The new approach is implemented in the tool General Purpose Petri net Simulator (GPenSIM). With this modular approach, subsystems can be modeled either as IO port-driven modules or as IO buffer-driven modules, yielding smaller yet powerful modules. In this paper: firstly, a brief literature review is presented on developing modular Petri net models. Secondly, modeling subsystems as modules are discussed. Finally, an application example is presented which involves a number of communicating agents. This work shows how modularization with GPenSIM can help modeling of large discrete systems where the communication infrastructure is dominated by communicating agents.

The paper presents a methodology of determination of technological allowances in the manufacturing process of semi-products of deciduous timber (oak). Soft modelling has been proposed for the determination of allowances for the variability of consecutive technological processes caused by the material, machine, operator, measurement system and random factors. The proposed methodology has been applied in the technological process of manufacturing the surface layer of a floor board (lamella). The method of identification of a soft model for the process of cutting timber into lamellas has been presented. Twenty soft models have been identified and recommendations have been made, based on experiments, for reducing the variability and centering the process. The proposed process modelling method facilitates economic estimation of allowances, controlling the manufacturing process and forecasting its future condition. As a result of the research, the manufacturing capacity of lamellas of acceptable quality has been increased by ca. 30%.

This paper attempts to solve the problem of searching minimum production order completion time variants by means of stochastic logical structures with all cost curve descent points and corresponding minimum-cost schedules. The analysis presented in this paper considers scheduling of unique and small batch production, predominantly to order, which accounts for changing requirements of the customer, the complexity and long production process makespan including its technical preparation. Scheduling of production order was performed by means of GAN networks and employed the concept of soft relations. The cost/time relation analysis is based on two-node network models using the cost curve. A new approach to scheduling under uncertainty is proposed and discussed. The problem is illustrated with an example.

Over the last decades, advances in diagnosis and tissue microsurgical reconstruction of soft tissues have modified the therapeutic approach to salivary gland cancers, but long term survival rates have increased only marginally. Due to the relatively low frequency of these tumors together with their diverse histopathological types, it is not easy to perform a prognosis assessment. Multivariate adaptative regression splines (MARS) is a data mining technique with a well-known ability to describe a response starting from a large number of predictors. In this work MARS was used for determining the prognosis of cancers of salivary glands using clinical and histological variables, as well as molecular markers. Here, we have generated four different models combining different sets of variables, with sensitivities and specificities that ranging from 95.45 to 100%. Specifically, one of these models which combined five clinical variables (Tumor size – T-, neck node metastasis – N-, distant metastasis – M-, age, and number of tumor recurrences) plus one molecular factor (gelatinase B -MMP-9-) showed a sensitivity and a specificity of 100%. Therefore, the MARS model was applied to the modelling of the influence of several clinical and molecular variables on the prognosis of salivary gland cancers with success. A self-organizing map (SOM) is a type of neural network what was used here to determine a prognostic model composed for four variables: N, M, number of recurrences and tumor type. The sensitivity of this model was that of 97%, and its specificity was that of 94.7%.

The objective of the present work was to create an artificial neural network model able to classify individuals suffering from bruxism in clenching and grinding patients according to the value of certain occlusal variables and other parameters. Patients suspected of bruxism represent a very heterogeneous group. Some require immediate treatment while others, with only minor disorders, may not need treatment at all.Artificial neural network (ANN) ensembles models were trained on with data from 325 bruxist patients examined at the Department of Prosthodontics and Occlusion (Craniomandibular Dysfunction Unit) of Oviedo University. The information retrieved from each patient included some occlusal variables and other information such as their gender and age. The models were analysed using Receiver Operating Characteristics (ROC) curve analysis, calibration assessments, inter- and intra-method variations. Effective odds ratios for the ANN ensembles were compared between each model.The ANN ensemble approach resulted in an area under the ROC curve of 86%. At 95% sensitivity the specificity was 84.1%, for the existence of 43.5% of bruxists clenching patients in the population of the study. This population corresponds to a grinding patients’ best predictive value of 97.2% and a clenching patients’ best predictive value of 89.5% both using the bagging method. The artificial neural network model obtained can distinguish between clenching and grinding patients requiring the analysis of a few variables and with a high rate of success.

The inclusion in workplaces of video display terminals has introduced multiple benefits in the organization of the work. Nevertheless, it also implies a series of risks for the health of the workers, since it can cause ocular and visual disorders, among others. In this work, a group of eye and vision-related problems associated to prolonged computer use (known as computer vision syndrome) are studied. The aim is to select the characteristics of the subject most relevant for the occurrence of this syndrome, and then, to develop a classification model for its prediction. The estimation of this problem is made by means of support vector machines for classification. This machine learning technique will be trained with the support of a genetic algorithm. This provides different patterns of parameters to the training of the support vector machine, improving its performance.The model performance is verified in terms of the area under the ROC curve, which leads to a model with high accuracy in the classification of the syndrome.

The objective of the present paper is to model the genetic influence in prostate cancer with Multivariate Adaptive Regression Splines (MARS) and Artificial Neural Networks (ANNs) techniques for classification. These models will be able to classify subjects that have cancer according to the values of the selected proteins from the genes selected with the models as most relevant. Subjects are selected as cases and controls from the MCC-Spain database and represent a heterogeneous group.Multivariate Adaptive Regression Splines models allow to select a set of the most valuables proteins from the database for modelling. These models were trained in 9 different degrees and chosen regarding its performance and complexity. Artificial neural networks models were trained on with data restricted to the most significant variables. The performance of both type of models were analysed in terms of the Area Under Curve (AUC) of the Receiver Operating Characteristics (ROC) curve. The ANN technique resulted in a model with AUC of 0.62006, while for MARS technique, the value was of 0.569312 in the best situation. Then, the artificial neural network model obtained can categorize if a patient suffer prostate cancer significantly better than MARS models and with high rate of success.

Nowadays, it is well-known that there are several genetic alterations that can be employed as genetic markers of prostate cancer. The use of pathways (gene sets) is one of the most promising areas of research in the cancer investigation. The aim of the present research is to study the influence of the pathways, with the help of models such as recursive partitioning method, to detect the single nucleotide polymorphism of relevance, and consequently the detection of prostate cancer. Data is retrieved from subjects of MCC-Spain database, and are selected as cases and controls, representing a heterogeneous group.With recursive partitioning method decision trees are built, which allow to prune off the splits that are supposed to be not of interest. Then, with the selected pathways, multivariate adaptive regression spline models are trained, and its performance is assessed in terms of the Area Under Curve (AUC) of the Receiver Operating Characteristics (ROC) curve.As results, with the help of performance tests, that would be useful for researchers that works with genetic datasets, a dimensional reduction and tuning of the parameters for the models is determined.In the case of our research, a total of 12 SNPs were found as the most relevant of the above mentioned database for the prostate cancer detection.

Background: Despite high success and survival rates of implant supported prosthesis therapy, biomechanical complications such as periimplant bone resorption continue to exist due to occlusal overloading.Purpose: To analyze and compare the influence of bone density, direction and distribution of the occlusal load and the design of three-unit implant supported prostheses.Materials and methods: Three dimensional finite element analysis study was developed to evaluate the influence on the periimplant bone stress distribution of three different designs of a 3-unit bridge supported by two implants (intermediate pontic, tilted implant, cantilever pontic), two different bone qualities (D3 and D4), and different loading directions (axial and non axial) and distributions (uniform and non uniform).Results: Bridge configuration with intermediate pontic and parallel implants presents the lowest periimplant stress, whereas the highest stress was found on the tilted implant bridge. Bone D3 is biomechanically more favourable than D4. Non axial and uniform loading conditions produce more periimplant stress.Conclusions: Distal cantilever and 45˚ convergent distal tilted implant are the second and third treatment options on a posterior three unit rehabilitation. Lightening occlusal contacts on the pontic and avoiding non axial loading, reduce periimplant bone stress. Bone D4 increases periimplant bone strain.

Colorectal cancer (CRC) has the third highest incidence in men, and the second highest in women worldwide. As it is known, both genetic and environmental factors play a role in colorectal cancer. So far, the most common way of studying genetic factors affecting CRC has been the SNP-SNP analysis. However, since these rarely act in an individualized way, it would be interesting to study them together. For that reasons, it is important to detect pathways or SNPs with a known relation which plays a role in this disease. In this study, we use Pathway of Distinction Analysis methodology (PoDA) in order to do it. PoDA is a novel bioinformatics tool that identifies significant pathways that could play an essential role in a specific disease based on genetics distance. Based on this method, we state that mitochondrial biogenesis pathway could be a good predictor pathway on colorectal cancer.

In this article, we present the use of sparse representation of signal and dictionary learning method for solving the problem of anomaly detection. The analyzed signal was presented as a set of correct ECG structures and outliers (characterizing different types of disorders). In the course of learning we used the modified Method of Optimal Directions (MOD) to find a dictionary that would reflect correct structures of an ECG signal. The dictionary found this way became a basis for sparse representation of the analyzed ECG signal. In the process of anomaly detection based on decomposition of the analyzed signal onto correct values and outliers, there was used a modified Alternating Minimization Algorithm (AMA). Performance of the proposed method was tested using a widely available database of ECG signals - MIT–BIH Arrhythmia Database. The obtained experimental results confirmed the effectiveness of the method of anomaly detection in the analysed ECG signals.

In some cases, big data bunches are in the form of Time Series (TS), where the occurrence of complex TS events are rarely presented. In this scenario, learning algorithms need to cope with the TS data balancing problem, which has been barely studied for TS datasets. This research addresses this issue, describing a very simple TS extension of the well-known SMOTE algorithm for balancing datasets. To validate the proposal, it is applied to a realistic dataset publicly available containing epilepsy-related TS. A study on the characteristics of the dataset before and after the performance of this TS balancing algorithm is performed, showing evidence on the requirements for the research on this topic, the energy efficiency of the algorithm and the TS generation process among them.

Causal sentences are a main part of the medical explanations, providing the causes of diseases or showing the effects of medical treatments. In medicine, causal association is frequently related to time restrictions. So, some drugs must be taken before or after meals, being ‘after’ and ‘before’ temporary constraints. Thus, we conjecture that frequently medical papers include time causal sentences. Causality involves a transfer of qualities from the cause to the effect, denoted by a directed arrow. An arrow connecting the node cause with the node effect is a causal graph. Causal graphs are an imagery way to show the causal dependencies that a sentence shows using plain text. In this paper, we will provide several programs to extract time causal sentences from medical Internet resources and to convert the obtained sentences in their equivalent causal graphs, providing an enlightening image of the relations that a text describes, showing the cause-effect links and the temporary constraints affecting their interpretation.

Monitoring pain and finding more efficient methods for analgesic administration during anaesthesia is a challenge that attracts the attention of both clinicians and engineers. This work focuses on the application of Machine Learning techniques to assist the clinicians in the administration of analgesic drug. The problem will consider patients undergoing general anaesthesia with intravenous drug infusion. The paper presents a preliminary study based on the use of the signal provided by an analgesia monitor, the Analgesia Nociception Index (ANI) signal. One aim of this research is studying the relation between ANI monitor and the changes in drug titration made by anaesthetist. Another aim is to propose an intelligent system that provides decisions on the drug infusion according to the ANI evolution. To do that, data from 15 patients undergoing cholecystectomy surgery were analysed. In order to establish the relationship between ANI and the analgesic, Machine Learning techniques have been introduced. After training different types of classifier and testing the results with cross validation method, it has been demonstrated that a relation between ANI and the administration of remifentanil can be found.

Risk prediction models for colorectal cancer play an important role to identify people at higher risk of developing this disease as well as the risk factors associated with it. Feature selection techniques help to improve the prediction model performance and to gain insight in the data itself. The assessment of the stability of feature selection/ranking algorithms becomes an important issue when the aim is to analyze the most relevant features. This work assesses several feature ranking algorithms in terms of performance and robustness for a set of risk prediction models. Experimental results demonstrate that stability and model performance should be studied jointly as RF turned out to be the most stable algorithm but outperformed by others in terms of model performance while SVM-wrapper and the Pearson correlation coefficient are moderately stable while achieving good model performance.

Public health in developed countries is heavily affected by pollution specially in highly populated areas. Amongst the pollutants with greatest impact in health, ozone is particularly addressed in this paper due to importance of its effect on cardiovascular and respiratory problems and their prevalence on developed societies. Local authorities are compelled to provide satisfactory predictions of ozone levels and thus the need of proper estimation tools rises. A data driven approach to prediction demands high quality data but those observations collected by weather stations usually fail to meet this requirement. This paper reports a new approach to robust ozone levels prediction by using an outlier detection technique in an innovative way. The aim is to assess the feasibility of using raw data without preprocessing in order to obtain similar or better results than with traditional outlier removal techniques. An experimental dataset from a location in Spain, Ponferrada, is used through an experimental stage in which such approach provides satisfactory results in a difficult case.

One of the main challenges in anesthesia is the proposal of safe and efficient methods to administer drugs to regulate the pain that the patient is sufffering during the surgical process. First steps towards this objective is the proposal of adequate indexes that correlate well with analgesia. One of the most promising index is ANI (Antinociception Index). This research focuses on the modelling of the ANI response in patients undergoing general anesthesia with intravenous drug infusion. The aim is to predict the ANI response in terms of the analgesic infusion rate. For this a model based on intelligent regression techniques is proposed. To create the model, it has been checked Artificial Neural Networks (ANN) and Support Vector Regression (SVR). Results were validated using data from patients in the operating room. The measured performance attest for the potential of the proposed technique.

This paper explores the benefit of using the PAELLA algorithm in an innovative way. The PAELLA algorithm was originally developed in the context of outlier detection and data cleaning. As a consequence, it is usually seen as a discriminant tool that categorizes observations into two groups: core observations and outliers. A new look at the information contained in its output provides ample opportunity in the context of data driven predictive models. The information contained in the occurrence vector is used through the experiments reported in a quest for finding how to take advantage of that information. The results obtained in each successive experiment guide the researcher to a sensible use case in which this information proves extremely useful: probabilistic sampling regression.

The task of retrieving a specific object from an image, which is similar to a query object is one of the critical applications in the computer vision domain. The existing methods fail to return similar objects when the region of interest is not specified correctly in a query image. Furthermore, when the feature vector is large, the retrieval from big collections is usually computationally expensive. In this paper, we propose an object retrieval method, which is based on the neural codes (activations) generated by the last inner-product layer of the Faster R-CNN network demonstrating that it can be used not only for object detection but for retrieval too. To evaluate the method, we have used a subset of ImageNet comprising of images related to indoor scenes, and to speed-up the retrieval, we first process all the images from the dataset and we save information (i.e. neural codes, objects present in the image, confidence scores and bounding box coordinates) corresponding to each detected object. Then, given a query image, the system detects the object present and retrieves its neural codes, which are then used to compute the cosine similarity against saved neural codes. We retrieved objects with high cosine similarity scores, and then we compared it with the results obtained using confidence scores. We showed that our approach takes only 0.534 s to retrieve all the 1454 objects in our test set.

This paper proposes two parallel algorithms to optimise a Fortran application that solves thermoelastic contact problems between three-dimensional solids using the Boundary Element Method. Parallel libraries like MPI are of great use when trying to minimise the execution time of numerical codes. Experiments carried out show the effectiveness of parallel programming and the study of the obtained results provides information on the main factors influencing that effectiveness. A reduction in the execution time of a 82.93% has been achieved.

The class of decimation-based sequence generators attempts to obtain an implicit non-linearity from the decimation process. In this work, it is shown that the output sequence of a well known member of this generator class, the shrinking generator, is composed of PN-sequences generated by Linear feedback Shift Registers. Furthermore, these PN-sequences are shifted versions of a unique sequence whose initial positions can be determined using discrete logarithms. Taking advantage of the linearity of the PN-sequences, a method of recovering the whole output sequence from a small number of intercepted bits is proposed. The algorithm is deterministic, always finds the cryptosystem key and is very adequate for parallelization. The basic ideas of this work can be generalized to other elements in the same class of sequence generators.

This paper proposes an adaptive database intrusion detection model that can be resistant to potential insider misuse with a limited number of data. The intrusion detection model can be adapted online using evolutionary reinforcement learning (ERL) which combines reinforcement learning and evolutionary learning. The model consists of two feedforward neural networks, a behavior network and an evaluation network. The behavior network detects the intrusion, and the evaluation network provides feedback to the detection of the behavior network. To find the optimal model, we encode the weights of the networks as an individual and produce populations of better individuals over generations. TPC-E scenario-based virtual query data were used for verification of the proposed model. Experimental results show that the detection performance improves as the proposed model learns the intrusion adaptively.

Relational databases contain information that must be protected such as personal information, the problem of intrusion detection of relational database is considered important. Also, the pattern of attacks evolves and it is difficult to grasp by rule-based method or general machine learning, so adaptive learning is needed. Learning classifier systems are system that combines supervised learning, reinforcement learning and evolutionary computation. It creates and updates classifiers according to data input. Learning classifier systems can learn adaptive because they generate and evaluate classifiers in real time. In this paper, we apply accuracy based learning classifier systems to relational database and confirm that adaptive learning is possible. Also, we confirmed their practical usability that they close to the best accuracy, though were not the best.

The preoccupation of the present work is an attempt to solve the problem of anomaly detection in network traffic by means of statistical models based on exponential smoothing. We used the generalized Holt-Winters model to detect possible fluctuations in network traffic, i.e. accidental fluctuations, trend and seasonal variations. The model parameters were estimated by means of the Hyndman-Khandakar algorithm. We chose the model parameters optimal values on the grounds of information criteria (AIC) which show a compromise between the consistency model and the size of its estimation error. In the proposed method, we used automatic forecasting on the basis of the estimated traffic model, which was further compared to the real variability of the analyzed network traffic in order to detect its abnormal behavior. The results of the performed experiments confirm efficiency of the proposed solution.

In the last years, cyber security of Industrial Control Systems (ICSs) has become an important issue due to the discovery of sophisticated malware that by attacking Critical Infrastructures, could cause catastrophic safety results. Researches have been developing countermeasures to enhance cyber security for pre-Internet era systems, which are extremely vulnerable to threats. This paper presents the potential opportunities that Software Defined Networking (SDN) provides for the security enhancement of Industrial Control Networks. SDN permits a high level of configuration of a network by the separation of control and data planes. In this work, we describe the affinities between SDN and ICSs and we discuss about implementation strategies.

The evolution of malicious software (malware) analysis tools provided controlled, isolated, and virtual environments to analyze malware samples. Several services are found on the Internet that provide to users automatic system to analyze malware samples, as VirusTotal, Jotti, or ClamAV, to name a few. Unfortunately, malware is currently incorporating techniques to recognize execution onto a virtual or sandbox environment. When analysis environment is detected, malware behave as a benign application or even show no activity. In this work, we present an empirical study and characterization of automatic public malware analysis services. In particular, we consider 26 different services. We also show a set of features that allow to easily fingerprint these services as analysis environments. Finally, we propose a method to mitigate fingerprinting.

The TOR Project allows the publication of content anonymously, which cause the proliferation of illegal material whose authorship is almost impossible to identify. In this paper, we present and make publicly available TOIC (TOr Image Categories), an image dataset which comprises five different illegal classes based on crawled TOR addresses. To classify those images we used Edge-SIFT features jointly with dense SIFT descriptors obtained from an “edge image” calculated with the Compass Operator. We demonstrate how a Bag of Visual Words model trained with the early fusion of dense SIFT and Edge-SIFT features can create an efficient model to detect and categorise illegal content in TOR network. Then, we estimated the radius for a complete dataset before the Edge-SIFT calculation, and we demonstrate that the classification performance is higher when the most salient edge information is extracted from the edges. We tested our proposal in both TOIC and in the public dataset Butterflies to prove the consistency of the method, obtaining an accuracy increase of 2.32 and 7.00 points respectively. We obtained with the Ideal Radius Selection an accuracy of 92.49% on TOIC dataset which makes this approach an attractive tool to detect and categorise illegal content in TOR network.

Password hashing functions are specialized cryptographic hash functions that have adjustable parameters with the aim of thwarting brute force or dictionary attacks over stolen password hash databases, even when specialized hardware or general purpose graphical processing units are employed.In this paper, we propose a password hashing function design based on the Advanced Encryption Standard block cipher in counter mode that has the advantage of being hardware accelerated on most modern computing platforms.We also analyze the performance, security and complexity characteristics of this algorithm and compare it to the well-known Scrypt password-hashing function with promising results.

We are presenting a highly-efficient, novel architecture (which we call FAST, or Forensic Analysis of Sensitive Traces) for high-performance big data forensics for heterogeneous systems (CPU and GPU-based). Our model uses a highly-compact storage format of the widely known Aho-Corasick algorithm [1], as well as a partial pruning mechanism to ensure the lowest possible memory footprint, while maximizing throughput performance. We are comparing our performance with classic methods used in data forensics and observe significant memory footprint improvements, as well as massive throughput improvements throughout all stages of big data processing.

Malware is one of the most dangerous threats that concerns cybersecurity. The main reasons for all of this are the development of Internet technology and Internet of Everything. Therefore, there are several mathematical models to simulate malware propagation and obtain countermeasures. These models are usually epidemic models based on ordinary differential equations. In this paper, we expose some of their deficiencies in order to improve the epidemic models. Moreover, we propose a new Susceptible-Carrier-Infectious-Recovered-Susceptible (SCIRS) model, which takes into account carrier devices. Finally, we demonstrate its global stability and study its dynamic behaviour through its basic reproductive number.

There are a great variety of specimens of malware: computer viruses, computer worms, trojans, etc. Nowadays, malware is one of the most important computer security problem and the source of great financial losses. Consequently, it is necessary to design tools that allow one to simulate the behavior of malware propagation. These tools are based on mathematical models and the great majority of them tackle the study of a particular type of malware called computer worms. Nevertheless, to the best of our knowledge, there are few models devoted to the study of the spreading of computer viruses. In this sense, the main goal of this work is to introduce a new mathematical model, based on cellular automata, to analyze the epidemic behavior of computer virus. Specifically, it is a SEIR (Susceptible-Exposed-Infectious-Recovered) model where the nodes of the network are divided into four compartments: susceptible, exposed, infected and recovered.

The distinctive security features of the Spanish electronic national identity card allow us to propose the usage of this cryptographic smart card in an authentication framework that can be used during the registration and login phases of certain internet services, including popular social networks. Using this mechanism with NFC-capable devices, the identity and age of the potential user can be determined, allowing or denying the access to the service based on that information.

Logs integration is one of the most challenging concerns in current security systems. Certainly, the accurate identification of security events requires to handle and merge highly heterogeneous sources of information. As a result, there is an urge to construct general codification and classification procedures to be applied on any type of security log. This work is focused on defining such a method using the so-called Normalised Compression Distance (NCD). NCD is parameter-free and can be applied to determine the distance between events expressed using strings. On the grounds of the NCD, we propose an anomaly-based procedure for identifying web attacks from web logs. Given a web query as stored in a security log, a NCD-based feature vector is created and classified using a Support Vector Machine (SVM). The method is tested using the CSIC-2010 dataset, and the results are analysed with respect to similar proposals.

In this work we show a deep review of lightweight random and pseudorandom number generators designed for constrained devices such as wireless sensor networks and RFID tags along with a study of Trifork pseudorandom number generator for constrained devices.

Cloud services provide a means to ease information storage and sharing. In the case of Small and Medium Enterprises, this represents a great opportunity to deploy platforms for data exchange without the high costs of traditional Information Technology solutions. Nonetheless, the adoption of the cloud implies a risk in terms of the security and privacy of the stored information assets. This risk is even more relevant when we opt for free cloud storage services. In this work we present Encrypted Cloud, a solution to endow users of free cloud storages with a mechanism to encrypt their information before uploading it. Since data sharing is a must for cloud users, Encrypted Cloud also enables the exchange of cryptographic keys in a secure way. Encrypted Cloud assumes a zero-trust context and correspondingly incorporates a thorough integrity verification protocol. Finally, Encrypted Cloud acts as a central entry point upon multiple free cloud storage services. This being the case, our tool enables users to allocate their resources in multiple cloud storages, and it also makes it possible to define a data redundancy policy.

In this paper we propose the practical application of multimodal conversational interfaces to develop advanced mobile learning applications. Our proposal integrates features of Android APIs on a modular architecture that emphasizes interaction management and context-awareness to build user-adapted, robust and maintainable (mobile learning) m-learning applications. We have applied the proposed framework to develop a mobile learning application that improves users’ interaction to solve practical educative exercises, promotes autonomous learning and self-assessment, personalizes the selection of recommended learning activities considering the student’s preferences and their previous uses, and allows the provision of immediate feedback by means of the automatic assessment.

Assessment is one of the most important practices for both teachers and students. In the classroom, the work of the students will be conditioned by the evaluation. In addition, students usually carry out the tasks with the only aim of improving their grades. The exclusive use of the lectures and their evaluation by means of an exam is an ineffective tool to achieve meaningful learning and the development of critical thinking. Currently, we have different methodologies that allow us to put these two concepts to practice, and also increase the interest and motivation of the students. The aim of this work is to s how the effectiveness of other evaluation methodologies.

Cybersecurity training is one of the relevant priorities for nations around the world in order to ensure a qualified, competitive and plentiful workforce in the years to come. It is necessary to promote the vocations of students towards cybersecurity studies to deal with the excess of uncovered labor demand with special attention to an under-represented group such as the female. To further this goal, our study focuses on a sample from among the group of students who have reached the postgraduate degree of MSc in Cybersecurity. Our methodology included in-person interviews and discussion groups and aims to identify the challenges/barriers that exist in the initial stages of career for young postgraduates as well as the skills and knowledge required for this group of young professionals.

The role of engineering is closely related to its role in society. An Industrial Engineer use to be involved in supervising the work of a team; in negotiations; and always may have special responsibilities to ensure that work is safe, and that work do not damage the environment. We have analyzed the specific case of ethics in industrial engineering undergraduate degrees at the University of Salamanca (Spain). In recent years, the White Books of qualifications for undergraduate degrees of the industrial branches, possesses a mention to deliver an optional subject related to ethics and the acquisition of the competence of an ethical and moral responsibility. In the case of the University of Salamanca, this competence is not included as a subject in the definition of the industrial engineering degrees. Ethical responsibility is included in the same group of transversal competencies as teamwork or critical reasoning. We have conducted a survey about ethics to different engineering students to get their feedback about the importance of ethical behaviour, the ethics and professional responsibility, or the necessity of be part of an official College of Industrial Engineers (after finishing university studies).

A new developed tool for PID learning is described on this work. The main contribution is the possibility to assist non-experimented users on the PID tuning task. For its implementation, knowledge engineering was used by the conceptual model creation and its next formalization on the described tool. Very good results have been achieved in general terms when it was validated with users without expertise on the control field. Both aims were achieved, the right learning of the traditional PID tuning by empirical methods and the assistance during tuning over systems.

The present study aims at understanding the challenges and identifying opportunities for improving teaching and learning in higher education institutions of diverse background. The case study used for data collection was the examinations and assignments developed along a course in project management for obtaining of the bachelor degree in Mechanical Engineering, together with a questionnaire. The course was taught in Spanish by the same teachers to students at University of León and at University of Xiangtan. Some of the students enrolled University of Xiangtan took the course at University of León. We assessed the students’ content knowledge during and after the course, competence and their perceptions about the overall course, content and instruction methods. Methodology approach consisting of master lectures, exercises, seminars and assignments. Students perceptions about the course were in general positive. Barriers to learning included student difficulties working in teams, course organization, limited student engagement in projects and previous knowledge. Language was the main difficulty for those students without a B2 Spanish certificate. This study aided in the identification of strategies to improve teaching and learning in highly diverse engineering environments.

This paper analyses the use of discussion forums by students that work together in the same project, and investigates their relationship with the final success of the project. Furthermore, the application of text mining techniques as a tool for identifying problems in the development of the learning experience is studied. Analyses were carried out using real data from students’ participation in project management and communication tools. The results show a strong positive correlation between the frequency of use of the discussion forums and the success of the project, as well as a low variability in the terminology used by the students. Despite the potential of text mining, the reduced use of vocabulary by the students makes the process of classifying messages into a complex problem.

According to the situated learning theory, meaningful learning cannot be considered outside an authentic context in which it normally occurs. Thus, this work presents a virtual learning environment that seeks to improve the quality of teaching and learning processes in higher education in the field of project management. To achieve these goals, we have adopted a student-centered approach that is supported by the use of tools of information and communication technologies. The proposed environment integrates different open source tools, such as a project and portfolio management software or a course management system, as well as specifically designed and developed tools to support administration, monitoring and assessment.