Case-Based Reasoning (CBR) is receiving increasedinterest for applications in medical decision support.Clinicians appreciate the fact that the system reasons withfull medical cases, symptoms, diagnosis, actions takenand outcomes. Also for experts it is often appreciated toget a second opinion. In the initial phase of a CBR systemthere are often a limited number of cases available whichreduces the performance of the system. If past cases aremissing or very sparse in some areas the accuracy isreduced. This paper presents a fuzzy rule-basedclassification scheme which is introduced into the CBRsystem to initiate the case library, providing improvedperformance in the stress diagnosis task. Theexperimental results showed that the CBR system usingthe enhanced case library can correctly classify 83% ofthe cases, whereas previously the correctness of theclassification was 61%. Consequently the proposedsystem has an improved performance with 22% in termsof accuracy. In terms of the discrepancy in classificationcompared to the expert, the goodness-of-fit value of thetest results is on average 87%. Thus by employing thefuzzy rule-based classification, the new hybrid system cangenerate artificial cases to enhance the case library.Furthermore, it can classify new problem cases previouslynot classified by the system.
Case-based reasoning systems for medical application are increasingly multi-purpose systems and also multi-modal, using a variety of different methods and techniques to meet the challenges from the medical domain. It this paper, some of the recent medical case-based reasoning systems are classified according to their functionality and development properties. It shows how a particular multi-purpose and multi-modal case-based reasoning system solved these challenges. For this a medical case-based reasoning system in the domain of psychophysiology is used.
Biofeedback is today a recognized treatment method for a number of physical and psychological problems. Experienced clinicians often achieve good results in these areas and their success largely builds on many years of experience and often thousands of treated patients. Unfortunately many of the areas where biofeedback is used are very complex, e.g. diagnosis and treatment of stress. Less experienced clinicians may even have difficulties to initially classify the patient correctly. Often there are only a few experts available to assist less experienced clinicians. To reduce this problem we propose a computer assisted biofeedback system helping in classification, parameter setting and biofeedback training. By adopting a case based approach in a computer-based biofeedback system, decision support can be offered to less experienced clinicians and provide a second opinion to experts. We explore how such a system may be designed and validate the approach in the area of stress where the system assists in the classification, parameter setting and finally in the training. In a case study we show that the case based biofeedback system outperforms novice clinicians based on a case library of cases authorized by an expert.
Biofeedback is a method gaining increased interest and showing good results for a number of physical and psychological problems. Biofeedback training is mostly guided by an experienced clinician and the results largely rely on the clinician's competence. In this paper we propose a three phase computer assisted sensor-based biofeedback decision support system assisting less experienced clinicians, acting as second opinion for experienced clinicians. The three phase CBR framework is deployed to classify a patient, estimate initial parameters and to make recommendations for biofeedback training by retrieving and comparing with previous similar cases in terms of features extracted. The three phases work independently from each other. Moreover, fuzzy techniques are incorporated into our CBR system to better accommodate uncertainty in clinicians reasoning as well as decision analysis. All parts in the proposed framework have been implemented and primarily validated in a prototypical system. The initial result shows how the three phases functioned with CBR technique to assist biofeedback training. Eventually the system enables the clinicians to allow a patient to train himself/herself unsupervised.
Intelligent analysis of heterogeneous data and information sources for efficient decision support presents an interesting yet challenging task in clinical envi-ronments. This is particularly the case in stress medicine where digital patient re-cords are becoming popular which contain not only lengthy time series measurements but also unstructured textual documents expressed in form of natural languages. This paper develops a hybrid case-based reasoning system for stress di-agnosis which is capable of coping with both numerical signals and textual data at the same time. The total case index consists of two sub-parts corresponding to signal and textual data respectively. For matching of cases on the signal aspect we present a fuzzy similarity matching metric to accommodate and tackle the imprecision and uncertainty in sensor measurements. Preliminary evaluations have revealed that this fuzzy matching algorithm leads to more accurate similarity estimates for improved case ranking and retrieval compared with traditional distance-based matching crite-ria. For evaluation of similarity on the textual dimension we propose an enhanced cosine matching function augmented with related domain knowledge. This is im-plemented by incorporating Wordnet and domain specific ontology into the textual case-based reasoning process for refining weights of terms according to available knowledge encoded therein. Such knowledge-based reasoning for matching of tex-tual cases has empirically shown its merit in improving both precision and recall of retrieved cases with our initial medical databases. Experts in the domain are very positive to our system and they deem that it will be a valuable tool to foster wide-spread experience reuse and transfer in the area of stress diagnosis and treatment.
Intelligent analysis of heterogeneous data and information sources for efficient decision support presents an interesting yet challenging task in clinical environments. This is particularly the case in stress medicine where digital patient records are becoming popular which contain not only lengthy time series measurements but also unstructured textual documents expressed in form of natural languages. This paper develops a hybrid case-based reasoning system for stress diagnosis which is capable of coping with both numerical signals and textual data at the same time. The total case index consists of two sub-parts corresponding to signal and textual data respectively. For matching of cases on the signal aspect we present a fuzzy similarity matching metric to accommodate and tackle the imprecision and uncertainty in sensor measurements. Preliminary evaluations have revealed that this fuzzy matching algorithm leads to more accurate similarity estimates for improved case ranking and retrieval compared with traditional distance-based matching criteria. For evaluation of similarity on the textual dimension we propose an enhanced cosine matching function augmented with related domain knowledge. This is implemented by incorporating Wordnet and domain specific ontology into the textual case-based reasoning process for refining weights of terms according to available knowledge encoded therein. Such knowledge-based reasoning for matching of textual cases has empirically shown its merit in improving both precision and recall of retrieved cases with our initial medical databases. Experts in the domain are very positive to our system and they deem that it will be a valuable tool to foster widespread experience reuse and transfer in the area of stress diagnosis and treatment.
Today, in our daily life we are subjected to a wide range of pressures. When the pressures exceed the extent that we are able to deal with then stress is trigged. High level of stress may cause serious health problems i.e. it reduces awareness of bodily symptoms. So, people may first notice it weeks or months later meanwhile the stress could cause more serious effect in the body and health. A difficult issue in stress management is to use biomedical sensor signals in the diagnosis and treatment of stress. This paper presents a case-based system that assists a clinician in diagnosis and treatment of stress. The system uses a finger temperature sensor and the variation in the finger temperature is one of the key features in the system. Several artificial intelligence techniques such as textual information retrieval, rule-based reasoning (RBR), and fuzzy logic have been combined together with case-based reasoning to enable more reliable and efficient diagnosis and treatment of stress. The performance has been validated implementing a research prototype and close collaboration with experts.
The amount of medical and industrial experience and knowledge is rapidly growing and it is almost impossible to be up to date with everything. The demand of decision support system (DSS) is especially important in domains where experience and knowledge grow rapidly. However, traditional approaches to DSS are not always easy to adapt to a flow of new experience and knowledge and may also show a limitation in areas with a weak domain theory. This chapter explores the functionalities of Case-Based Reasoning (CBR) to facilitate experience reuse both in clinical and in industrial decision making tasks. Examples from the field of stress medicine and condition monitoring in industrial robots are presented here to demonstrate that the same approach assists both for clinical applications as well as for decision support for engineers. In the both domains, DSS deals with sensor signal data and integrate other artificial intelligence techniques into the CBR system to enhance the performance in a number of different aspects. Textual information retrieval, Rule-based Reasoning (RBR), and fuzzy logic are combined together with CBR to offer decision support to clinicians for a more reliable and efficient management of stress. Agent technology and wavelet transformations are applied with CBR to diagnose audible faults on industrial robots and to package such a system. The performance of the CBR systems have been validated and have shown to be useful in solving such problems in both of these domains.
Experience is one of the most valuable assets technicians and engineer have and may have been collected during many years and both from successful solutions as well as from very costly mistakes. Unfortunately industry rarely uses a systematic approach for experience reuse. This may be caused by the lack of efficient tools facilitating experience distribution and reuse. We propose a case-based approach and tool to facilitate experience reuse more systematically in industry. It is important that such a tool allows the technicians to give the problem case in a flexible way to increase acceptance and use. The proposed tool enables more structured handling of experience and is flexible and can be adapted to different situations and problems. The user is able to input text in a structured way and possibly in combination with other numeric or symbolic features. The system is able to identify and retrieve relevant similar experiences for reuse.
Industry has to adjust quickly to changes in their surroundings, for example reducing staff during recession and increasing staff when the market demands it. These factors may cause rapid loss of experience, collected during many years, or require experienced staff to spend considerable resources in training new staff, instead of focusing on production. This is recognised as very costly for companies and organisations today and also reduces competitiveness and productivity. Condition Monitoring, diagnostics and selection of efficient preventive or corrective actions is a task that often requires a high degree of expertise. This expertise is often gained through sometimes very expensive mistakes and can take many years to acquire leading to a few skilled experts. When they are not available due to changes in staff or retirements the company often faces serious problems that may be very expensive, e.g. leading to a reduced productivity.
If some deviation occurs in a machine, a fault report is often written; an engineer makes a diagnosis and may order spare parts to repair the machine. Fault report, spare parts, required time and statistics on performance after repair are often stored in different databases but so far not systematically reused. In this paper we present a Case-Based experience sharing system that enables reuse of experience in a more efficient way compared with what is mostly practiced in industry today. The system uses Case-Based-Reasoning (CBR) and limited Natural Language Processing. An important aspect of the experience management tool is that it is user-friendly and web-based to promote efficient experience sharing. The system should be able to handle both experiences that are only in house as well as sharing experience with other industries when there is no conflicting interest. Such a CBR based tool enables efficient experience gathering, management and reuse in production industries. The tool will facilitate the users with an interactive environment to communicate with each other for sharing their experiences. Depend on the user; the security level of the system will be varied to share knowledge among the collaborating companies.
The system identifies the most relevant experiences to assess and resolve the current situation. The experience is stored and retrieved as a case in the collaborative space where experience from various companies may have been stored under many years. Reusing experience and avoiding expensive mistakes will increase the participating companies' competitiveness and also transfer valuable experience to their employees. One of the benefits is also the opportunity and facility to identify people with similar tasks and problems at different companies and enable them to share their experience, e.g. if a technician has solved a similar problem recently and is in the near, the most efficient solution may be to call the expert and ask for assistance. In future, one may access this tool through his/her mobile device via wireless or mobile communications using Global Positioning System, GPS, enables the system to suggest experts nearby, willing and able to share the knowledge and quickly assist in resolve the problem.
A common problem for autonomous vehicles is to define a coherent round boundary of unstructured roads. To solve this problem an evolutionary approach has been evaluated, by using a modified ant optimization algorithm to define a coherent road edge along the unstructured road in night conditions. The work presented in this paper involved pre-processing, perfecting the edges in an autonomous fashion and developing an algorithm to find the best candidates of starting positions for the ant colonies. All together these efforts enable ant colony optimization (ACO) to perform successfully in this application scenario. The experiment results show that the best paths well followed the edges and that the mid-points between the paths stayed centered on the road.
PID controllers are still popular in a wide range of engineering practices due to their simplicity and robustness. Traditional design of a PID controller needs manual setting of its parameters in advance. This paper proposes a new method for online tuning of PID controllers based on hybridized neural membrane computing. A neural network is employed to adaptively determine the proper values of the PID parameters in terms of evolving situations/stages in the control process. Further the learning of the neural network is performed based on a membrane algorithm, which is used to locate the weights of the network to optimize the control performance. The effectiveness of the proposed method has been demonstrated by the preliminary results from simulation tests.
Case-based reasoning systems for medical application are increasingly applied to meet the challenges from the medical domain. This paper looks at the state of the art in case-based reasoning and some systems are classified in this respect. A survey is performed based on the recent publications and research projects in CBR in medicine. Also, the survey is based on e-mail questionnaire to the authorsâ to complete the missing property information. Some clear trends in recent projects/systems have been identified such as most of the systems are multi-modal, using a variety of different methods and techniques to serve multipurpose i.e. address more than one task.
The Health Sciences are, nowadays, one of the major application areas for case-based reasoning (CBR). The paper presents a survey of recent medical CBR systems based on a literature review and an e-mail questionnaire sent to the corresponding authors of the papers where these systems are presented. Some clear trends have been identified, such as multipurpose systems: more than half of the current medical CBR systems address more than one task. Research on CBR in the area is growing, but most of the systems are still prototypes and not available on the market as commercial products. However, many of the projects/systems are intended to be commercialized.
Increased exposure to stress may cause serious health problems leading to long term sick leave if undiagnosed and untreated. The practice amongst clinicians' to use a standardized procedure measuring blood pressure, ECG, finger temperature, breathing speed etc. to make a reliable diagnosis of stress and stress sensitivity is increasing. But even with these measurements it is still difficult to diagnose due to large individual variations. A computer-based system as a second option for the assessment and diagnosis of individual stress level is valuable in this domain.
A combined approach based on a calibration phase and case-based reasoning is proposed exploiting data from finger temperature sensor readings from 24 individuals. In calibration phase, a standard clinical procedure with six different steps helps to establish a person's stress profile and set up a number of individual parameters. When acquiring a new case, patients are also asked to provide a fuzzy evaluation on how reliable was the procedure to define the case itself. Such a reliability "level" could be used to further discriminate among similar cases. The system extracts key features from the signal and classifies individual sensitivity to stress. These features are stored into a case library and similarity measurements are taken to assess the degrees of matching and create a ranked list containing the most similar cases retrieved by using the nearest-neighbor algorithm.
A current case (CC) is compared with two other stored cases (C_92 and C_115) in the case library. The global similarity between the case CC and case C_92 is 67% and case CC and case C_115 is 80% shown by the system. So the case C_115 has ranked higher than the case C_92 and is more similar to current case CC. If necessary, the solution for the best matching case can be revised by the clinician to fit the new patient. The current problem with confirmed solution is then retained as a new case and added to the case library for future use.
The system allows us to utilize previous experience and at the same time diagnose stress along with a stress sensitivity profile. This information enables the clinician to make a more informed decision of treatment plan for the patients. Such a system may also be used to actively notify a person's stress levels even in the home environment.
Stress diagnosis based on finger temperature signals is receiving increasing interest in the psycho-physiological domain. However, in practice, it is difficult and tedious for a clinician and particularly less experienced clinicians to understand, interpret and analyze complex, lengthy sequential measurements in order to make a diagnosis and treatment plan. The paper presents a case-based decision support system to assist clinicians in performing such tasks. Case-based reasoning is applied as the main methodology to facilitate experience reuse and decision explanation by retrieving previous similar temperature profiles. Further fuzzy techniques are also employed and incorporated into the case-based reasoning system to handle vagueness, uncertainty inherently existing in clinicians reasoning as well as imprecision of feature values. Thirty nine time series from 24 patients have been used to evaluate the approach (matching algorithms) and an expert has ranked and estimated similarity. On average goodness-of-fit for the fuzzy matching algorithm is 90% in ranking and 81% in similarity estimation which shows a level of performance close to an experienced expert. Therefore, we have suggested that a fuzzy matching algorithm in combination with case-based reasoning is a valuable approach in domains where the fuzzy matching model similarity and case preference is consistent with the views of domain expert. This combination is also valuable where domain experts are aware that the crisp values they use have a possibility distribution that can be estimated by the expert and is used when experienced experts reason about similarity. This is the case in the psycho-physiological domain and experienced experts can estimate this distribution of feature values and use them in their reasoning and explanation process.
Increased exposure to stress may cause health problems. An experi-enced clinician is able to diagnose a person's stress level based on sensor read-ings. Large individual variations and absence of general rules make it difficult to diagnose stress and the risk of stress-related health problems. A decision sup-port system providing clinicians with a second opinion would be valuable. We propose a novel solution combining case-based reasoning and fuzzy logic along with a calibration phase to diagnose individual stress. During calibration a num-ber of individual parameters are established. The system also considers the feedback from the patient on how well the test was performed. The system uses fuzzy logic to incorporating the imprecise characteristics of the domain. The cases are also used for the individual treatment process and transfer experience between clinicians. The validation of the approach is based on close collabora-tion with experts and measurements from 24 persons used as reference.
Diagnosing stress is difficult even for experts due to large individual variations. Clinician's use today manual test procedures where they measure blood pressure, ECG, finger temperature and breathing speed during a number of exercises. An experienced clinician makes diagnosis on different readings shown in a computer screen. There are only very few experts who are able to diagnose and predict stress-related problems. In this paper we have proposed a combined approach based on a calibration phase and case-based reasoning to provide assistance in diagnosing stress, using data from the finger temperature sensor readings. The calibration phase helps to establish a number of individual parameters. The system uses a case-based reasoning approach and also feedback on how well the patient succeeded with the different test, used for giving similar cases reliability estimates.
The increasing use of digital patient records in hospital saves both time and reduces risks wrong treatments caused by lack of information. Digital patient records also enable efficient spread and transfer of experience gained from diagnosis and treatment of individual patient. This is today mostly manual (speaking with col-leagues) and rarely aided by computerized system. Most of the content in patient re-cords is semi-structured textual information. In this paper we propose a hybrid tex-tual case-based reasoning system promoting experience reuse based on structured or unstructured patient records, case-based reasoning and similarity measurement based on cosine similarity metric improved by a domain specific ontology and the nearest neighbor method. Not only new cases are learned, hospital staff can also add comments to existing cases and the approach enables prototypical cases.
In the medical literature there are a number of physiological reactions related to cognitive activities. Psychosocial and psychophysiological stress is such activities reflected in physiological reactions. Stress related symptoms are highly individual, but decreased hands temperature is the common for most individuals. A clinician learns with experience how to interpret the different symptoms but there is no adaptive diagnostic system for diagnosing stress. Decision support systems (DSS) diagnosing stress would be valuable both for junior clinicians and as second opinion for experts. Due to the large individual variations and no general set of rules, DSS are difficult to build for this task. The proposed solution combines a calibration phase with case-based reason¬ing approach and fuzzification of cases. During the calibration phase a number of individual parameters and case specific fuzzy membership functions are es-tablishes. This case-based approach may help the clinician to make a diagnosis, classification and treatment plan. The case may also be used to follow the treat-ment progress. This may be done using the proposed system. Initial tests show promising results. The individual cases including calibration and fuzzy mem-bership functions may also be used in an autonomous system in home environ-ment for treatment programs for individuals often under high stress.
Today, everyday life for many people contain many situations that may trigger stress or result in an individual living on an increased stress level under long time. High level of stress may cause serious health problems. It is known that respiratory rate is an important factor and can be used in diagnosis and biofeedback training, but available measurement of respiratory rate are not especially suitable for home and office use. The aim of this project is to develop a portable sensor system that can measure the stress level, during everyday situations e.g. at home and in work environment and can help the person to change the behaviour and decrease the stress level. The sensor explored is a finger temperature sensor. Clinical studies show that finger temperature, in general, decreases with stress; however this change pattern shows large individual variations. Diagnosing stress level from the finger temperature is difficult even for clinical experts. Therefore a computer-based stress diagnosis system is important. In this system, case-based reasoning and fuzzy logic have been applied to assists in stress diagnosis and biofeedback treatment utilizing the finger temperature sensor signal. An evaluation of the system with an expert in stress diagnosis shows promising result.
The complexity of modern lifestyle and society brings many advantages but also causes increased levels of stress for many people. It is known that increased exposure to stress may cause serious health problems if undiagnosed and untreated and a report from the Swedish government estimates that 1/3 of all long term sick leave is stress related. One of the physiological parameters for quantifying stress levels is the finger temperature that helps the clinician in diagnosis and treatment of stress. However, in practice, the complex and varying nature of signals makes it difficult and tedious to interpret and analyze the lengthy sequential measurements. A computer based system diagnosing stress would be valuable both for clinicians and for treatment. Since stress diagnosis has a week domain theory and there are large individual variations, Case-Based Reasoning is proposed as the main methodology. Feature extraction methods abstracting the original signals without losing key features are investigated. A fuzzy technique is also incorporated into the system to perform matching between the features derived from signals to better accommodate vagueness, uncertainty often present in clinical reasoning Validation of the approach is based on close collaboration with experts and measurements from twenty four persons. The system formulates a new problem case with 17 extracted features from the fifteen minutes (1800 samples) of biomedical sensor data. Thirty nine time series from twenty four persons have been used to evaluate the approach (matching algorithms) in which the system shows a level of performance close to an experienced expert. The system can be used as an expert for a less experienced clinician, as a second option for an experienced clinician or for treatment in home environment.
Federated learning (FL) is an emerging and privacy-preserving machine learning technique that is shown to be increasingly important in the digital age. The two challenging issues for FL are: (1) communication overhead between clients and the server, and (2) volatile distribution of training data such as class imbalance. The paper aims to tackle these two challenges with the proposal of a federated fuzzy learning algorithm (FFLA) that can be used for data-based construction of fuzzy classification models in a distributed setting. The proposed learning algorithm is fast and highly cheap in communication by requiring only two rounds of interplay between the server and clients. Moreover, FFLA is empowered with an an imbalance adaptation mechanism so that it remains robust against heterogeneous distributions of data and class imbalance. The efficacy of the proposed learning method has been verified by the simulation tests made on a set of balanced and imbalanced benchmark data sets.
Experienced staffs acquire their experience during many years of practice, and sometimes also through expensive mistakes. This knowledge is often lost when technicians retire, or if companies need to downsize during periods of reduced sale. When scaling up production, new staff requires training and may repeat similar mistakes. Another issue that may be costly is when monitoring systems repeatedly give false alarms, causing expensive loss of production capacity and resulting in technicians losing trust in the systems and in worst case, switch them off. If monitoring systems could learn from previous experience for both correct and false alarms, the reliability and trust in the monitoring systems would increase. Moreover, connecting alarms to either equipment taking automatic actions or recommend actions based on the current situations and previous experience would be valuable.
An engineer repeating the same task a second time is often able to perform the task in 1/3 of the time it took at the first time. Most corrective and preventive actions for a particular machine type have been carried out before. This past experience holds a large potential for time savings, predictability and reduced risk if an efficient experience transfer can be accomplished. But building large complex support system is not always the ideal way. We propose instead localized intelligent agents, able to either autonomously perform the necessary actions or aid a human in the decision making process by providing the necessary information needed to make an informed and validated decision.
This paper discusses the role and integration of knowledge discovery (KD) in case-based reasoning (CBR) systems. The general view is that KD is complementary to the task of knowledge retaining and it can be treated as a separate process outside the traditional CBR cycle. Unlike knowledge retaining that is mostly related to case-specific experience, KD aims at the elicitation of new knowledge that is more general and valuable for improving the different CBR substeps. KD for CBR is exemplified by a real application scenario in medicine in which time series of patterns are to be analyzed and classified. As single pattern cannot convey sufficient information in the application, sequences of patterns are more adequate. Hence it is advantageous if sequences of patterns and their co-occurrence with categories can be discovered. Evaluation with cases containing series classified into a number of categories and injected with indicator sequences shows that the approach is able to identify these key sequences. In a clinical application and a case library that is representative of the real world, these key sequences would improve the classification ability and may spawn clinical research to explain the co-occurrence between certain sequences and classes.
This paper addresses the issue of discovering key sequences from time series data for pattern classification. The aim is to find from a symbolic database all sequences that are both indicative and non-redundant. A sequence as such is called a key sequence in the paper. In order to solve this problem we first we establish criteria to evaluate sequences in terms of the measures of evaluation base and discriminating power. The main idea is to accept those sequences appearing frequently and possessing high co-occurrences with consequents as indicative ones. Then a sequence search algorithm is proposed to locate indicative sequences in the search space. Nodes encountered during the search procedure are handled appropriately to enable completeness of the search results while removing redundancy. We also show that the key sequences identified can later be utilized as strong evidences in probabilistic reasoning to determine to which class a new time series most probably belongs.
Coping with time series cases is becoming an important issue in case based reasoning. This paper develops a knowledge discovery approach to discovering significant sequences for depicting symbolic time series cases. The input is a case library containing time series cases consisting of consecutive discrete patterns. The proposed approach is able to find from the given case library all qualified sequences that are non-redundant and indicative. A sequence as such is termed as a key sequence. It is shown that the key sequences discovered are highly usable in case characterization to capture important properties while ignoring random trivialities. The main idea is to transform an original (lengthy) time series into a more concise representation in terms of the detected occurrences of key sequences. Three alternate ways to develop case indexes based on key sequences are suggested. These indexes are simply vectors of numbers that are easily usable when matching two time series cases for case retrieval.
In many industrial and medical diagnosis problems it is essential to investigate time series measurements collected to recognize existing or potential faults/diseases. Today this is usually done manually by humans. However the lengthy and complex nature of signals in practice often makes it a tedious and hard task to analyze and interpret available data properly even by experts with rich experiences. The incorporation of intelligent data analysis method such as case-based reasoning is showing strong benefit in offering decision support to technicians and clinicians for more reliable and efficient judgments. This chapter addresses a general framework enabling more compact and efficient representation of practical time series cases capturing the most important characteristics while ignoring irrelevant trivialities. Our aim is to extract a set of qualitative, interpretable features from original, and usually real-valued time series data. These features should on one hand convey significant information to human experts enabling potential discoveries/findings and on the other hand facilitate much simplified case indexing and imilarity matching in case-based reasoning. The road map to achieve this goal consists of two subsequent stages. In the first stage it is tasked to transform the time series of real numbers into a symbolic series by temporal abstraction or symbolic approximation. A few different methods are available at this stage and they are introduced in this chapter. Then in the second stage we use knowledge discovery method to identify key sequences from the transformed symbolic series in terms of their cooccurrences with certain classes. Such key sequences are valuable in providing concise and important features to characterize dynamic properties of the original time series signals. Four alternative ways to index time series cases using discovered key sequences are discussed in this chapter.
Monitoring, quality control and diagnosis is a large cost for production industry. Studies have estimated that the total cost of maintenance in Sweden is 20 billion Euros and the amount spent on maintenance in Europe is around 1500 billion Euros per year. The key to efficient maintenance is monitoring and quality control. Much of this work is today still manual and based on experienced technicians. Today large amounts of data are collected in the production industry but only a fragment of this data is used. Much of the monitoring data from sensors are used for quality control and maintenance which is still interpreted manually or a system monitoring if a threshold value is passed in order to give an alert. More elaborate use of the data, information and experience is rare. Using methods and techniques from artificial intelligence for experience reuse enables more informed actions based reducing accidents, mistakes and costs to mention some benefits. Building up and sharing experience is the key to intelligent monitoring and diagnostics acting as decision support. Intelligent Monitoring Agents are going beyond decision support since they also have communication skill and are able to make decisions on their own. Keywords: diagnostic systems, monitoring, artificial intelligence, agent based architecture.
Differences in data size per class, also known as imbalanced data distribution, have become a common problem affecting data quality. Big Data scenarios pose a new challenge to traditional imbalanced classification algorithms, since they are not prepared to work with such amount of data. Split data strategies and lack of data in the minority class due to the use of MapReduce paradigm have posed new challenges for tackling the imbalance between classes in Big Data scenarios. Ensembles have been shown to be able to successfully address imbalanced data problems. Smart Data refers to data of enough quality to achieve high-performance models. The combination of ensembles and Smart Data, achieved through Big Data preprocessing, should be a great synergy. In this paper, we propose a novel Smart Data driven Decision Trees Ensemble methodology for addressing the imbalanced classification problem in Big Data domains, namely SD_DeTE methodology. This methodology is based on the learning of different decision trees using distributed quality data for the ensemble process. This quality data is achieved by fusing random discretization, principal components analysis, and clustering-based random oversampling for obtaining different Smart Data versions of the original data. Experiments carried out in 21 binary adapted datasets have shown that our methodology outperforms random forest.
Digital Twins (DTs) serve as the backbone of Industry 4.0, offering virtual representations of actual systems, enabling accurate simulations, analysis, and control. These representations help predict system behaviour, facilitate multiple real-time tests, and reduce risks and costs while identifying optimization areas. DTs meld cyber and physical realms, accelerating the design and modelling of sustainable innovations. Despite their potential, the complexity of DTs presents challenges in their industrial application. We continue here the development of our approach to build an adaptable and trustable framework for building and operating DT systems - A Digital Twin Framework for Dynamic and Robust Distributed Systems (D-RODS). D-RODS aims to address the challenges above, aiming to advance industrial digitalization and targeting areas like system efficiency, incorporating AI and verification techniques with formal support. We employ existing large-usage tools to illustrate the approach in development based on a synthetic adaptable use case.
Digital Twins (DTs) are virtual representations of physical products in many dimensions, such as geometry and behaviour. As a backbone of Industry 4.0, DTs help interpret and even predict the behaviour of physical processes, provide a virtual testbed for maintenance and upgrade, and enable automatic decision-making supported by artificial intelligence. Despite the promising future, challenges exist, such as the absence of a framework that facilitates the development and application of DTs in industrial contexts. We propose a service-oriented architecture (SOA) DT framework for dynamic and robust distributed systems. The framework contains two types of services. One includes the services provided to the users and is supported by an orchestration mechanism to ensure a quality of service (QoS). The other one refers to the common functions of all DTs. Further, we describe the DT-based decision-making enabled by our QoS-oriented learning of the framework and a Hoare-logic-based verification of QoS.
Most real-time control systems are often accompanied with various changes such as variations of working load and changes of the environment. Hence it is necessary to perform real-time process modeling so that the model can adjust itself in runtime to maintain high accuracy of states under control. This paper considers process model represented as a deep recurrent neural network. We propose a new hybridized learning method for online updating the weights of such recurrent neural networks by exploiting both fast convergence of Kalman filter and stable search of the Backpropagation through time algorithm. Several experiments were made to show that the proposed learning method has fast convergence, high accuracy and good adaptivity. It can not only achieve high modeling accuracy for a static process but also quickly adapt to changes of characteristics in a time -varying process.
This paper addresses the feature selection problems in the setting of online learning of data streams. Typically this setting imposes restrictions on computational resources (memory, processing) as well as storage capacity, since instances of streaming data arrive with high speed and with no possibility to store data for later offline processing. Feature selection can be particularly beneficial here to selectively process parts of the data by reducing the data dimensionality. However selecting a subset of features may lead to permanently ruling out the possibilities of using discarded dimensions. This will cause a problem in the cases of feature drift in which data importance on individual dimensions changes with time. This paper proposes a new method of online feature selection to deal with drifting features in non-stationary data streams. The core of the proposed method lies in deep reconstruction networks that are continuously updated with incoming data instances. These networks can be used to not only detect the point of change with feature drift but also dynamically rank the importance of features for feature selection in an online manner. The efficacy of our work has been demonstrated by the results of experiments based on the MNIST database.
Case Based Reasoning is a method for solving problems based on previous experience. In some cases large numbers of cases are available representing the experience to draw from. Dealing with large numbers of potentially noisy cases introduces challenges to do with storage capacity and similarity matching but can also lead to sub-optimal solutions. This paper proposes an algorithm for generating a reduced set of highly representative constructed cases, to use as the Smart Case Base in replacement of the original large training data set. The goal of using the reduced set of cases is to address the problems linked to the original Case Base size as well as to maintain or to improve upon generalization accuracy for problem solving. The algorithm makes use of membrane clustering, an evolutionary computing approach for organizing data in large search spaces. The performance of the algorithm has been evaluated using public benchmark data sets.
Motor Imagery (MI) classification from neural activity is thought to represent valuable information that can be provided as real-time feedback during rehabilitation after for example a stroke. Previous studies have suggested that MI induces partly subject-specific EEG activation patterns, suggesting that individualized classification models should be created. However, due to fatigue of the user, only a limited number of samples can be recorded and, for EEG recordings, each sample is often composed of a large number of features. This combination leads to an undesirable input data set for classification. In order to overcome this constraint, we propose a new methodology to create and select features from the EEG signal in two steps. First, the input data is divided into different windows to reduce the cardinality of the input. Secondly, a Hierarchical Genetic Algorithm is used to select relevant features using a novel fitness function which combines the data reduction with a correlation feature selection measure. The methodology has been tested on EEG oscillatory activity recorded from 6 healthy volunteers while they performed an MI task. Results have successfully proven that a classification above 75% can be obtained in a restrictive amount of time (0.02 s), reducing the number of features by almost 90%.
Accurate system modelling is an important prerequisite for optimized process control in modern industrial scenarios. The task of parameter identification for a model can be considered as an optimization problem of searching for a set of continuous parameters to minimize the discrepancy between the model outputs and true output values. Differential Evolution (DE), as a class of population-based and global search algorithms, has strong potential to be employed here to solve this problem. Nevertheless, the performance of DE is rather sensitive to its two running parameters: scaling factor and crossover rate. Improper setting of these two parameters may cause weak performance of DE in real applications. This paper presents a new adaptive algorithm for DE, which does not require good parameter values to be specified by users in advance. Our new algorithm is established by integration of greedy search into the original DE algorithm. Greedy search is conducted repeatedly during the running of DE to reach better parameter assignments in the neighborhood. We have applied our adaptive DE algorithm for process model identification in a Furnace Optimized Control System (FOCS). The experiment results revealed that our adaptive DE algorithm yielded process models that estimated temperatures inside a furnace more precisely than those produced by using the original DE algorithm.
Differential evolution (DE), as a class of biologically inspired and meta-heuristic techniques, has attained increasing popularity in solving many real world optimization problems. However, DE is not always successful. It can easily get stuck in a local optimum or an undesired stagnation condition. This paper proposes a new DE algorithm Differential Evolution with Alopex-Based Local Search (DEALS), for enhancing DE performance. Alopex uses local correlations between changes in individual parameters and changes in function values to estimate the gradient of the landscape. It also contains the idea of simulated annealing that uses temperature to control the probability of move directions during the search process. The results from experiments demonstrate that the use of Alopex as local search in DE brings substantial performance improvement over the standard DE algorithm. The proposed DEALS algorithm has also been shown to be strongly competitive (best rank) against several other DE variants with local search.
Differential evolution (DE) presents a class of evolutionary and meta-heuristic techniques that have been applied successfully to solve many real-world problems. However, the performance of DE is significantly influenced by its control parameters such as scaling factor and crossover probability. This paper proposes a new adaptive DE algorithm by greedy adjustment of the control parameters during the running of DE. The basic idea is to perform greedy search for better parameter assignments in successive learning periods in the whole evolutionary process. Within each learning period, the current parameter assignment and its neighboring assignments are tested (used) in a number of times to acquire a reliable assessment of their suitability in the stochastic environment with DE operations. Subsequently the current assignment is updated with the best candidate identified from the neighborhood and the search then moves on to the next learning period. This greedy parameter adjustment method has been incorporated into basic DE, leading to a new DE algorithm termed as Greedy Adaptive Differential Evolution (GADE). GADE has been tested on 25 benchmark functions in comparison with five other DE variants. The results of evaluation demonstrate that GADE is strongly competitive: it obtained the best rank among the counterparts in terms of the summation of relative errors across the benchmark functions with a high dimensionality.
Differential evolution (DE) is a population-based metaheuristic algorithm that has been proved powerful in solving a wide range of real-parameter optimization tasks. However, the selection of the mutation strategy and control parameters in DE is problem dependent, and inappropriate specification of them will lead to poor performance of the algorithm such as slow convergence and early stagnation in a local optimum. This paper proposes a new method termed as Joint Adaptation of Parameters in DE (JAPDE). The key idea lies in dynamically updating the selection probabilities for a complete set of pairs of parameter generating functions based on feedback information acquired during the search by DE. Further, for mutation strategy adaptation, the Rank-Based Adaptation (RAM) method is utilized to facilitate the learning of multiple probability distributions, each of which corresponds to an interval of fitness ranks of individuals in the population. The coupling of RAM with JAPDE results in the new RAM-JAPDE algorithm that enables simultaneous adaptation of the selection probabilities for pairs of control parameters and mutation strategies in DE. The merit of RAM-JAPDE has been evaluated on the benchmark test suit proposed in CEC2014 in comparison to many well-known DE algorithms. The results of experiments demonstrate that the proposed RAM-JAPDE algorithm outperforms or is competitive to the other related DE variants that perform mutation strategy and control parameter adaptation, respectively.
Differential Evolution represents a class of evolutionary algorithms that are highly competitive for solving numerical optimization problems. In a Differential Evolution algorithm, there are a few alternative mutation strategies, which may lead to good or a bad performance depending on the property of the problem. A new mutation strategy, called DE/Alopex/1, is proposed in this paper. This mutation strategy distinguishes itself from other mutation strategies in that it uses the fitness values of the individuals in the population in order to calculate the probabilities of move directions. The performance of DE/Alopex/1 has been evaluated on the benchmark suite from CEC2013. The results of the experiments show that DE/Alopex/1 outperforms some state-of-the-art mutation strategies. © 2017 IEEE.
Differential evolution (DE) presents a class of evolutionary computing techniques that appear effective to handle real parameter optimization tasks in many practical applications. However, the performance of DE is not always perfect to ensure fast convergence to the global optimum. It can easily get stagnation resulting in low precision of acquired results or even failure. This paper proposes a new memetic DE algorithm by incorporating Eager Random Search (ERS) to enhance the performance of a basic DE algorithm. ERS is a local search method that is eager to replace the current solution by a better candidate in the neighborhood. Three concrete local search strategies for ERS are further introduced and discussed, leading to variants of the proposed memetic DE algorithm. In addition, only a small subset of randomly selected variables is used in each step of the local search for randomly deciding the next trial solution. The results of tests on a set of benchmark problems have demonstrated that the hybridization of DE with Eager Random Search can substantially augment DE algorithms to find better or more precise solutions while not requiring extra computing resources.
This paper proposes a memetic computing algorithm by incorporating Eager Random Search (ERS) into differential evolution (DE) to enhance its search ability. ERS is a local search method that is eager to move to a position that is identified as better than the current one without considering other opportunities. Forsaking optimality of moves in ERS is advantageous to increase the randomness and diversity of search for avoiding premature convergence. Three concrete local search strategies within ERS are introduced and discussed, leading to variants of the proposed memetic DE algorithm. The results of evaluations on a set of benchmark problems have demonstrated that the integration of DE with Eager Random Search can improve the performance of pure DE algorithms while not incurring extra computing expenses.
Differential evolution has many mutation strategies which are problem dependent. Some Adaptive Differential Evolution techniques have been proposed tackling this problem. But therein all individuals are treated equally without taking into account how good these solutions are. In this paper, a new method called Ranked-based Mutation Adaptation (RAM) is proposed, which takes into consideration the ranking of an individual in the whole population. This method will assign different probabilities of choosing different mutation strategies to different groups in which the population is divided. RAM has been integrated into several well-known adaptive differential evolution algorithms and its performance has been tested on the benchmark suit proposed in CEC2014. The experimental results shows the use of RAM can produce generally better quality solutions than the original adaptive algorithms.
Differential evolution (DE) is a very attractive evolutionary and meta-heuristic technique to solve many optimization problems in various real-world scenarios. However, the proper setting of control parameters of DE is highly dependent on the problem to solve as well as on the different stages of the search process. This paper proposes a new greedy adaptation method for dynamic adjustment of mutation factor and crossover rate in DE. The proposed method is based on the idea of greedy search to find better parameter assignment in the neighborhood of a current candidate. Our work emphasizes reliable evaluation of candidates via applying a candidate with a number of times in the search process. As our purpose is not merely to increase the success rate (the survival of more trial solutions) but also to accelerate the speed of fitness improvement, we suggest a new metric termed as progress rate to access the quality of candidates in support of the greedy search. This greedy parameter adaptation method has been incorporated into basic DE, leading to a new DE algorithm called Greedy Adaptive Differential Evolution (GADE). GADE was tested on 25 benchmark functions in comparison with five other DE variants. The results of evaluation demonstrate that GADE is strongly competitive: it obtains the best ranking among the counterparts in terms of the summation of relative errors across the benchmark functions.
Differential evolution (DE) is one competitive form of evolutionary algorithms. It heavily relies on mutating solutions using scaled differences of randomly selected individuals from the population to create new solutions. The choice of a proper mutation strategy is important for the success of an DE algorithm. This paper presents an empirical investigation to examine and compare the different mutation strategies for global optimization problems. Both solution quality and computational expense of DE variants were evaluated with experiments conducted on a set of benchmark problems. The results of such comparative study would offer valuable insight and information to develop optimal or adaptive mutation strategies for future DE researches and applications.
Differential Evolution is a stochastic and metaheuristic technique that has been proved powerful for solving real valued optimization problems in high dimensional spaces. However, Differential Evolution does not guarantee to con verge to the global optimum and it is easily to become trapped in a local optimum. In this paper, we aim to enhance Differential Evolution with Random Local Search to increase its ability to avoid local optimum. The proposed new algorithm is called Differential Evolution with Random Local Search (DERLS). The advantage of Random Local Search used in DERLS is that it is simple and fast in computation. The results of experiments have demonstrated that our DERLS algorithm can bring appreciable improvement for the acquired solutions in difficult optimization problems.