Search

Population pharmacokinetic analysis is used to estimate pharmacokinetic parameters and their variability from concentration data. Due to data sparseness issues, available datasets often do not allow the estimation of all parameters of the suitable model. The PRIOR subroutine in NONMEM supports the estimation of some or all parameters with values from previous models, as an alternative to fixing them or adding data to the dataset. From a literature review, the best practices were compiled to provide a practical guidance for the use of the PRIOR subroutine in NONMEM. Thirty-three articles reported the use of the PRIOR subroutine in NONMEM, mostly in special populations. This approach allowed fast, stable and satisfying modelling. The guidance provides general advice on how to select the most appropriate reference model when there are several previous models available, and to implement and weight the selected parameter values in the PRIOR function. On the model built with PRIOR, the similarity of estimates with the ones of the reference model and the sensitivity of the model to the PRIOR values should be checked. Covariates could be implemented a priori (from the reference model) or a posteriori, only on parameters estimated without prior (search for new covariates). Graphic abstract Electronic supplementary material The online version of this article (10.1007/s10928-020-09695-z) contains supplementary material, which is available to authorized users.

Realizing the accurate prediction of data flow is an important and challenging problem in industrial automation. However, due to the diversity of data types, it is difficult for traditional time series prediction models to have good prediction effects on different types of data. To improve the versatility and accuracy of the model, this paper proposes a novel hybrid time-series prediction model based on recursive empirical mode decomposition (REMD) and long short-term memory (LSTM). In REMD-LSTM, we first propose a new REMD to overcome the marginal effects and mode confusion problems in traditional decomposition methods. Then use REMD to decompose the data stream into multiple in intrinsic modal functions (IMF). After that, LSTM is used to predict each IMF subsequence separately and obtain the corresponding prediction results. Finally, the true prediction value of the input data is obtained by accumulating the prediction results of all IMF subsequences. The final experimental results show that the prediction accuracy of our proposed model is improved by more than 20% compared with the LSTM algorithm. In addition, the model has the highest prediction accuracy on all different types of data sets. This fully shows the model proposed in this paper has a greater advantage in prediction accuracy and versatility than the state-of-the-art models. The data used in the experiment can be downloaded from this website: https://github.com/Yang-Yun726/REMD-LSTM .

Most clinicians have come to imagine that scientific evidence has levels, and at the top of the pyramid are randomized clinical trials (RCTs) and meta-analyses [12]. But we all have read some poorly designed randomized trials, and sometimes a humble case series can substantially influence thought about a technique or implant, particularly if it describes failures of treatment. Nevertheless, as the Centre for Evidence-Based Medicine analogizes [6], if clinical evidence were stored in boxes marked “randomized trial,” “cohort study,” “case series,” and the like, and we needed to answer an important question, we might first open the box stamped “randomized trial” to see if there is anything inside. All the same, there are some important kinds of questions that RCTs and meta-analyses will never answer. Surgical RCTs generally are small to medium-sized, and as such, are tooled to assess efficacy more than safety. Randomized trials have almost no ability to detect rare events, evaluate resource utilization in normal practice, or compare any but the most common complications. Likewise, premature failures of commonly used interventions probably will not be identified first by RCTs, but rather by studies of other designs. To assess real-world and important endpoints like these, we will need to look elsewhere for evidence. Registry studies and analyses drawn from large-system databases represent two robust approaches to answering these kinds of questions. Clinical Orthopaedics and Related Research® is proud to support the excellent, thought-provoking work being done using the world’s registries. We publish it frequently [10, 11], and we have promoted it in Spotlight commentaries and interviews on these pages [13]. CORR® will publish selected proceedings from the most-recent meeting of the International Society of Arthroplasty Registries this summer, and we will cover the important role registries play in clinical research in an editorial in the coming months. We also believe that important, pressing questions can be answered efficiently and accurately using large databases, including the Nationwide Inpatient Sample (NIS), National Surgical Quality Improvement Program (NSQIP), National Hospital Discharge Survey (NHDS), National Trauma Databank (NTDB), and Medicare administrative databases, among others. In fact, there has been something of a bloom of orthopaedic papers from these kinds of sources lately [2–4]. With sample sizes that run from thousands to millions of patients, these studies allow us to tackle questions that likely will never be answered any other way. To make best use of this material, readers need to know just what kinds of questions can be asked of databases, how the same question may elicit different answers from different databases, and how slipshod or carelessly presented work from databases can mislead us. Finally, authors should know how editors for CORR® evaluate this relatively new type of research. We believe studies drawn from these databases best serve clinicians when they focus on what they are uniquely suited to do: Correlate less-common adverse events with modifiable and previously unidentified risk factors [1], identify adjustable provider- or hospital-level variables associated with readmissions or complications [7], and compare resource utilization for common interventions across diverse geographic regions or practice settings [5]. As with any other study design, the more novel the questions (and the more unexpected the answers), the more important these studies are. They do not make such good reading when they tell us what we already know. Readers also should realize that the same question asked of different databases may result in different answers [2, 4]. The most obvious reason for this is that each database surveys a specific population and may gather different data in different ways. Some databases, like the NIS, are inpatient-only samples: The moment a patient leaves the hospital, (s)he no longer is followed. As a result, studies drawn from inpatient-only samples that ask about complications that occur both before and after discharge—such as surgical site infection—will badly underestimate event rates. By contrast, prospective surgical registry-type databases (such as NSQIP) follow patents for 30 days or longer after surgery, and so are better tooled to find problems that occur after discharge. What do they lose in the exchange? Sample size. Prospective registries are smaller. There are other differences as well. Some registries focus on particular types of patients. The NTDB, for instance, includes patients with higher-energy injuries. Clearly, matching the study population to the question being asked is crucial. Additionally, each database uses different data elements. Many databases, including NIS, are built from ICD-9 codes. Since these codes’ primary use is billing, they have lower sensitivity to answer the clinical questions in which most readers are interested [8, 9]. By contrast, other databases—including NSQIP and parts of the NTDB—use chart-abstracted data, which offer more of the details clinicians seek. Readers should also recognize that missing data limits all databases to some degree, and any study drawn from them should comment on the degree to which this might influence its findings. We believe the most important failing of orthopaedic studies from these sources is that these databases generally do not capture enough patient-level diagnostic information or patient-reported outcomes data. Efforts are underway to address many of these limitations. But at present, questions calling for that level of detail probably are better studied using single or multicenter designs that capture more complete patient-level data of this sort. Notwithstanding the shortcomings of large-database studies, we at CORR® believe that they have an important role to play. But since these studies are not always easy reading, there must be a payoff for the reader who toughs it out—at the very least, a finding that changes the way the reader thinks about an important problem or treats his or her patients. Going forward, CORR® will consider a study that draws from a large database if it either presents a genuinely counterintuitive descriptive finding, or provides a specific suggestion to improve clinical care, practice management, or public policy. We will publish papers meeting one or both of those criteria when they are methodologically robust. As with the all of the important studies we publish in CORR®, we will continue to make every attempt to accompany these studies with CORR® Insights commentaries or Editor’s Spotlight features.

In the transition from paper to electronic workflow, the University of Colorado Health System’s implementation of a new electronic health record system (EHR) forced all clinical groups to reevaluate their practices including the infrastructure surrounding clinical trials. Radiological imaging is an important piece of many clinical trials and requires a high level of consistency and standardization. With EHR implementation, paper orders were manually transcribed into the EHR, digitizing an inefficient work flow. A team of schedulers, radiologists, technologists, research personnel, and EHR analysts worked together to optimize the EHR to accommodate the needs of research imaging protocols. The transition to electronic workflow posed several problems: (1) there needed to be effective communication throughout the imaging process from scheduling to radiologist interpretation. (2) The exam ordering process needed to be automated to allow scheduling of specific research studies on specific equipment. (3) The billing process needed to be controlled to accommodate radiologists already supported by grants. (4) There needed to be functionality allowing exams to finalize automatically skipping the PACS and interpretation process. (5) There needed to be a way to alert radiologists that a specialized research interpretation was needed on a given exam. These issues were resolved through the optimization of the “visit type,” allowing a high-level control of an exam at the time of scheduling. Additionally, we added columns and fields to work queues displaying grant identification numbers. The build solutions we implemented reduced the mistakes made and increased imaging quality and compliance.

Graphical abstract In spite of a rapid growth of data processing software, that has allowed for a huge advancement in many fields of chemistry, some research issues still remain problematic. A standard example of a troublesome challenge is the analysis of multi-component mixtures. The classical approach to such a problem consists of separating each component from a sample and performing individual measurements. The advent of computers, however, gave rise to a relatively new domain of data processing – chemometry – focused on decomposing signal recorded for the sample rather than the sample itself. Regrettably, still a very few chemometric methods are practically used in everyday laboratory routines. The Authors believe that a brief ‘user-friendly’ guide-like article on several ‘flagship’ algorithms of chemometrics may, at least partly, stimulate an increased interest in the use of these techniques among researchers specializing in many fields of chemistry. In the paper, five different techniques of factor analysis are used for the analysis of a three-component system of fluorophores. These algorithms, applied on the excitation-emission spectra, recorded for the ‘unknown’ mixture, allowed to unambiguously determine its composition without the need for physical separation of the components. An example of using chemometric methods for physical chemistry research is also provided. For each presented technique of the data analysis, a short description of its theoretical background followed by an example of its practical performance is given. In addition, the Reader is supplemented with a basic information on matrix algebra, detailed experimental ‘recipes’, reference specialist literature and ready-to-use MATLAB codes. Supplementary Information The online version contains supplementary material available at 10.1007/s10895-021-02753-w.

In this paper, we discuss the problems of arbitrage detection, which is known as change point detection in statistics. There are some classical methods for change point detection, such as the cumulative sum (CUSUM) procedure. However, when utilizing CUSUM, we must be sure about the model of the data before detecting. We introduce a new method to detect the change points by using Hilbert–Huang transformation (HHT) to devise a new algorithm. This new method (called the HHT test in this paper) has the advantage in that no model assumptions are required. Moreover, in some cases, the HHT test performs better than the CUSUM test, and has better simulation results. In the end, an empirical study of the volatility change based on the S&P 500 is also given for illustration.

COVID-19 is an ongoing pandemic disease. To make more accurate diagnosis on COVID-19 than existing approaches, this paper proposed a novel method combining DenseNet and optimization of transfer learning setting (OTLS) strategy. Preprocessing was used to enhance, crop, and resize the collected chest CT images. Data augmentation method was used to increase the size of training set. A composite learning factor (CLF) was employed which assigned different learning factor to three types of layers: frozen layers, middle layers, and new layers. Meanwhile, the OTLS strategy was proposed. Finally, precomputation method was utilized to reduce RAM storage and accelerate the algorithm. We observed that optimization setting "201-IV" can achieve the best performance by proposed OTLS strategy. The sensitivity, specificity, precision, and accuracy of our proposed method were 96.35 ± 1.07, 96.25 ± 1.16, 96.29 ± 1.11, and 96.30 ± 0.56, respectively. The proposed DenseNet-OTLS method achieved better performances than state-of-the-art approaches in diagnosing COVID-19.

Data quality improvement has become a critical issue for many companies and organizations because poor data quality degrades organizational performance whereas improved data quality results in cost saving and customer satisfaction. Activities such as identify and removing "duplicate" database records from a single database, and correlating records from different databases that identify the same real world "entity" are used routinely to improve data quality. Due to huge databases having several hundred millions to several billions records, and continuously growing, efficient techniques and algorithms are needed. One way to speed up the processing is to adopt a two-step approach, where potential candidate records are grouped together in step one. After that each group is further processed and analyzed in step two. The so called "transitive closure problem" is a formal formulation of what needs to be done in step one. This paper introduces a record grouping problem called transitive closure, and proposes algorithms to solve the transitive closure problem. The paper also reports on the empirical study of the proposed algorithms and comments on their implementations.

Since the spatial pattern of forest pest occurrence is determined by biological characteristics and habitat conditions, this paper introduced construction of a cellular automaton model combined with Markov model to predicate the forest pest occurrence. Rules of the model includes the cell states rules, neighborhood rules and transition rules which are defined according to the factors from stand conditions, stand structures, climate and the influence of the factors on the state conversion. Coding for the model is also part of the implementations of the model. The participants were designed including attributes and operations of participants expressed with a UML diagram. Finally, the scale issues on forest pest occurrence prediction, of which the core are the prediction of element size and time interval, are partly discussed in this paper.

Network convergence across different access technologies holds a promise of enabling ubiquitous service availability but faces several technical challenges. With anticipated proliferation of multimode IP devices, the optimal selection of a service delivery network among multiple IP based wireless access alternatives is one of the important issues that is actively studied and discussed in several standardization forums. Use of multi attribute decision making (MADM) algorithms has been proposed in the past for network selection decisions in a heterogeneous wireless network environment. A direct comparison of these algorithms is difficult as this would require the use of another MADM algorithm. A better approach instead is to ascertain the appropriateness of the algorithm to the problem space. This paper provides the basis for evaluating the appropriateness of MADM algorithms for network selection. It analyzes the use of MADM algorithms such as TOPSIS, ELECTRE and GRA for network selection and argues that GRA provides the best approach in scenarios where the utilities of some of the attributes are non-monotonic. The paper proposes a novel stepwise approach for GRA that uses multiple reference networks and explains its working with network selection scenarios.

With the rapid growth of software size and complexity, how to understand and maintain a software system has become a great challenge in its life-cycle, especially from the global perspective. The paper mainly focuses on the comprehension problem of database-based software systems. A framework for visualizing and analyzing table dependency structure of a DB-based application is proposed, and its corresponding prototype tool named DBViewer is implemented. It mainly includes two main functions, i.e., table dependency visualization in database and structure feature metrics. The key methods of building this tool are proposed and some characters of table dependency relations in database are also found via metrics in this paper. Some experiments are performed on five DB-based software systems, and some preliminary rules have been found: (1) the degree distribution of table dependency graph roughly conforms to power law, (2) the average path length of the max connected subgraph is small, and (3) the clustering effect is not so evident and the coefficients are all less than 0.05 in the max connected sub-graph of all table dependency relations.

Bayesian network is a powerful tool to represent and deal with uncertain knowledge. There exists much uncertainty in crop or animal disease. The construction of Bayesian network need much data and knowledge. But when data is scarce, some methods should be adopted to construct an effective Bayesian network. This paper introduces a disease diagnosis model based on Bayesian network, which is two-layered and obeys noisy-or assumption. Based on the two-layered structure, the relationship between nodes is obtained by domain knowledge. Based on the noisy-model, the conditional probability table is elicited by three methods, which are parameter learning, domain expert and the existing certainty factor model. In order to implement this model, a Bayesian network tool is developed. Finally, an example about cow disease diagnosis was implemented, which proved that the model discussed in this paper is an effective tool for some simple disease diagnosis in crop or animal field.

This paper presents an approach for mining fuzzy Association Rules (ARs) relating the properties of composite items, i.e. items that each feature a number of values derived from a common schema. We partition the values associated to properties into fuzzy sets in order to apply fuzzy Association Rule Mining (ARM). This paper describes the process of deriving the fuzzy sets from the properties associated to composite items and a unique Composite Fuzzy Association Rule Mining (CFARM) algorithm founded on the certainty factor interestingness measure to extract fuzzy association rules. The paper demonstrates the potential of composite fuzzy property ARs, and that a more succinct set of property ARs can be produced using the proposed approach than that generated using a non-fuzzy method.

Many researchers and organizations are interested in creating a mechanism capable of automatically predicting software defects. In the last years, machine learning techniques have been used in several researches with this goal. Many recent researches use data originated from NASA (National Aeronautics and Space Administration) IV&V (Independent Verification & Validation) Facility Metrics Data Program (MDP). We have recently applied a constructive neural network (RBF-DDA) for this task, yet MLP neural networks were not investigated using these data. We have observed that these data sets contain inconsistent patterns, that is, patterns with the same input vector belonging to diAEerent classes. This paper has two main objectives, (i) to propose a modified version of RBF-DDA, named RBF-eDDA (RBF trained with enhanced Dynamic Decay Adjustment algorithm), which tackles inconsistent patterns, and (ii) to compare RBF-eDDA and MLP neural networks in software defects prediction. The simulations reported in this paper show that RBF-eDDA is able to correctly handle inconsistent patterns and that it obtains results comparable to those of MLP in the NASA data sets.

A web-based multi-models expert system called DCDDS is presented in this paper, which developed for diagnosis of dairy cow diseases through the symptoms submitted by users on web. As it is accepted that the inference engine and the relevant knowledge representation are the crucial part of diagnosis expert system, which limits its application and popularization in animal disease diagnosis. To break the limit and raise accuracy, this paper compares and appraises the existed systems and presents a solution that contains three models—Case-based reasoning (CBR), Subjective Bayesian theory and D-S evidential theory. Accordingly a knowledge representation method which can support the three different models is also designed. Up to the complicacy of the group of symptoms users acquired, they can choose which of the three models should be adopted to meet the best resolve. The performance of the proposed system was evaluated by an application to the field of dairy cow disease diagnosis using a real example of dairy cow diseases. The result indicates that the new methods have improved the inference procedures of the expert systems, and have showed that the new architecture has some advantage over the conventional architectures of expert systems on both efficiency and accuracy.

As enterprises increase their electronic communication in business activities through network, the security of enterprise information system based on EC becomes crucial for enterprises. This paper proposes a fuzzy integrated security evaluation method based on man-computer combined data collection and fuzzy expert evaluation in Delphi method. The method could reduce the subjectivity of expert evaluation and alleviate the difficulty of data collection and makes possible a better combination of qualitative and quantitative evaluations. Firstly, the evaluation factor and hierarchy structure of security evaluation is constructed. Secondly, according to data collected and the evaluation comment of each expert, the subjection degree matrix is constructed. Finally, a new concept of “degree of assurance” is presented for the quantificational evaluation of enterprise information system security based on. In this paper the study of the case shows that the method can be easily used and its results conform to the actual situation.

More and more literatures have researched the application of data mining technology in customer segmentation, and achieved sound effects. One of the key purposes of customer segmentation is customer retention. But the application of single data mining technology mentioned in previous literatures is unable to identify customer churn trend for adopting different actions on customer retention. This paper focus on constructs a integrated data mining and survival analysis model to segment customers into heterogeneous group by their survival probability (churn trend) and help enterprises adopting appropriate actions to retain profitable customers according to each segment’s churn trend. This model contains two components. Firstly, using data mining clustering arithmetic cluster customers into heterogeneous clusters according to their survival characters. Secondly, using survival analysis predicting each cluster’s survival/hazard function to identify their churn trend and test the validity of clustering for getting the correct customer segmentation. This model proposed by this paper was applied in a dataset from one biggest china telecommunications company. This paper also suggests some propositions for further research.

Considering the information transmission system of agriculture M-informatization, pertinence and practicability are critical factors, but the current system lack these factors in China. Based on the case study of Agriculture Space Time of China Unicom and Agricultural ICT of China Mobile, the segmentation transmission approach is presented in this paper. To make the current information transmission system more pertinent and practical, we adopt segmentation transmission approach in the critical nodes of the closed loop information transmission system, which includes produce, gathering, analysis & process, supplying, sending, distribution, receiving, application and feedback. According to the key attribute dimensionality of users, this approach employs intelligent IMS (Information Matching System) to match the content of data dictionary with the individual needs of users, and analyze the static and dynamic attributes of users’ datum and the content of service by using the high-powered rule-engine. This data dictionary is based on the scientific multilayer-integrated data dictionary concerning farm products. Adopting the approach presented in this paper in current agriculture M-informationization system can successfully transmit all kinds of segmentation agriculture information to the right agriculture users.

This paper proposes a new formalized definition of generalized association rule based on Multidimensional data. The algorithms named BorderLHSs and GenerateLHSs-Rule are designed for generating generalized association rule from multi-level frequent item sets based on Multidimensional Data. Experiment shows that the algorithms proposed in this paper are more efficiency, generate less redundant rules and have good performance in flexibility, scalability and complexity.

The costing and data mining of logistics costs will become increasingly important to all firms seeking competitive advantages. Activity Based Costing (ABC) is considered as the optimized and most promising method of costing and controlling logistics cost now, and logistics cost data mining based on ABC is playing a very important role in business management. The paper firstly analyses the costing principle of ABC for logistics projects, and according to the basic principles of ABC and business management, the paper puts forward the basically technical route of logistics project costing and data mining based on ABC for business management and decision-making, moreover, the concrete costing process and model of applying ABC are deduced, then the further application forms of data mining based on ABC are summarized and elaborated for business management and decision-making.

In Information Retrieval (IR), the user’s input query conditions usually are not detailed enough, so the satisfactory query results can not be brought back. Query expansion of IR can help to solve this problem. However, the common query expansion in IR cannot get steady retrieval results. In this paper, we propose and implement query expansion method which combines domain ontology with the frequent of terms. Ontology is used to describe domain knowledge; logic reasoner and the frequency of terms are used to choose fitting expansion words. By this way, higher recall and precise can be gotten as user’ query results. Experimental results show that compared with the results of common query expansion, the method described in this paper can get statistically significant improvement in recall and precise combination.

Environmental pollution control relies heavily on human expert judgment sup- ported by historical data and scientific models. Telemonitoring, by networks of heterogeneous sensor arrays, provides the opportunity for data mining models to be constructed from the historical data to supplement human expertise. This paper reports some progress made in the TELEMAC project by data mining. TELEMAC is concerned with enhancing the efficacy of anaerobic digestion in potentially unstable digesters. In the laboratory using full instrumentation it is possible to derive a good description of the digester state. With data mining it is possible to identify some constraints on sensor choice. This paper examines this data mining work from the perspective of a three layer Grid architecture to see what implications and requirements arise that could benefit the exercise of expert judgment. After placing the specific TELEMAC situation in a generic Grids context, we present a classification approach to attributes for metadata and indicate some examples of model resource discovery.

In the process of data integration using ontologies it is important to manage data from external data sources in the same way as data stored in the Knowledge Base. In previous papers [1], [2] the way of inference from data stored in the Knowledge Base, using Knowledge Cartography idea has been presented. However, this solution requires loading all data to the Knowledge Base. The solution presented in this paper shows how the Knowledge Cartography can be used to infer from data stored in external data sources, without loading them to the Knowledge Base. The presented solution is to enrich each data source with an additional layer that allows managing data using signatures. The paper additionally describes the results of tests comparing times of inference when data are loaded to the Knowledge Base and when data are fetched on demand.

This book contains a selection of the best papers given at the Advanced Computer Systems Conference in October 2006 in Poland. The book is partitioned into four topical areas: The Artificial Intelligence section includes papers on Humatronics, Fuzzy Logic, Robotics, Cell, Genetic and Probabilistic Algorithms, Incomplete Data and Similarity Relations, and Fast Neural Networks. The Computer Security and Safety section covers a wide range of issues, from Error Detection in Block Ciphers to Visual Cryptography Methods. The Image Analysis, Graphics and Biometrics section deals with the extraction of meaningful information from images as well as methods for uniquely recognizing humans based upon intrinsic physical or behavioral traits. The Computer Simulation and Data Analysis section concerns computer programs that attempt to simulate an abstract model of a particular system and transform data with the aim of extracting useful information and facilitating conclusions.

This paper suggests several estimation guidelines for the choice of a suitable machine learning technique for software development effort estimation. Initially, the paper presents a review of relevant published studies, pointing out pros and cons of specific machine learning methods. The techniques considered are Association Rules, Classification and Regression Trees, Bayesian Belief Networks, Neural Networks and Clustering, and they are compared in terms of accuracy, comprehensibility, applicability, causality and sensitivity. Finally the study proposes guidelines for choosing the appropriate technique, based on the size of the training data and the desirable features of the extracted estimation model.

In this paper, we discuss the visualization of multidimensional data. A well-known procedure for mapping data from a high-dimensional space onto a lower-dimensional one is Sammon’s mapping. The algorithm is oriented to minimize the projection error. We investigate an unsupervised backpropagation algorithm to train a multilayer feed-forward neural network (SAMANN) to perform the Sammon’s nonlinear projection. Sammon mapping has a disadvantage. It lacks generalization, which means that new points cannot be added to the obtained map without recalculating it. The SAMANN network offers the generalization ability of projecting new data, which is not present in the original Sammon’s projection algorithm. Retraining of the network when the new data points appear has been analyzed in this paper.

This paper will discuss the motivations and methods for collecting quantitative data about free, libre and open source (FLOSS) software projects. The paper also describes the current state of the art in collecting this data, and some of the problems with this process. Finally, the paper outlines the challenges data miners should look forward to when trying to improve the usefulness of their quantitative data streams.

This paper discusses platform independent conceptual modeling of a knowledge intensive application, focusing on the use of knowledge-based systems (KBS) in the context of model-driven engineering. An extension to the Unified Modeling Language (UML) for knowledge modeling is presented based on the profiling extension mechanism of UML. The UML profile discussed in this paper has been successfully captured in a Meta-Object-Facility (MOF) based UML tool — the executable Modeling Framework (XMF). The example is that of modeling a knowledge-based system for the Ulcer Clinical Practical Guidelines (CPG) Recommendations. It demonstrates the use of the profile, with the prototype system implemented in the Java Expert System Shell (JESS).

One problem domain of meta search is to combine and improve the precision of ranking results from various search systems. This paper describes a rank aggregation model that incorporates text analysis measure with existing rank-based method, e.g. Best Rank and Borda Rank, to aggregate search results from various search systems. This approach provides means to normalize the differences of rank methodology used by different search systems, justifying the potential of using contents analysis to improve the results relevancy in meta search. In this paper, we fully describe our approach on text normalization for meta search and present our rationality of using two rank-based methods in our model. We then evaluate and benchmark the performance of our model based on user judgment on results relevancy. Our experiment results show that when text analysis factor is taken into account, the results outperform the rank-based methods alone. This shows the potential of our model to complement current rank aggregation methods used in meta search.

In this paper we present a novel data mining technique, known as Post Sequential Patterns Mining, which can be used to discover Structural Patterns. A Structural Pattern is a new pattern, which is composed of sequential patterns, branch patterns or iterative patterns. Sequential patterns mining plays an essential role in many areas and substantial research has been conducted on their analysis and applications. In our previous work [12], we used a simple but efficient Sequential Patterns Graph (SPG) to model the sequential patterns. The task to discover hidden Structural Pattern is based on our previous work and sequential patterns mining, conveniently named Post Sequential Patterns Mining. In this paper, in addition to stating this new mining problem, we define patterns such as branch pattern, iterative pattern, structural pattern, and concentrate on finding concurrent branch pattern. Concurrent branch pattern is thus one of the main forms of structural pattern and will play an important role in event-based data modelling.

This paper presents exploratory results on how a data-mining-based tool can be used to enhance the quality of decision-making in a Virtual Enterprise environment. The developed tool is based on the Clustering mining method and implements the K-Means algorithm. The algorithm is explained, its utilization in the proposed model is introduced and the implementation results are presented and stressed in the end of the paper.

Intrusion alert correlation is the process to identify high-level attack scenarios by reasoning about low-level alerts raised by intrusion detection systems (IDSs). The efficiency of intrusion alert correlation is critical in enabling interactive analysis of intrusion alerts as well as prompt responses to attacks. This paper presents an experimental study aimed at adapting main memory index structures (e.g., T Trees, Linear Hashing) and database query optimization techniques (e.g., nested loop join, sort join) for efficient correlation of intensive alerts. By taking advantage of the characteristics of the alert correlation process, this paper presents three techniques named hyper-alert container, two-level index, and sort correlation. This paper then reports a series of experiments designed to evaluate the effectiveness of these techniques. These experiments demonstrate that (1) hyper-alert containers improve the efficiency of order-preserving index structures (e.g., T Trees), with which an insertion operation involves search, (2) two-level index improves the efficiency of all index structures, (3) a two-level index structure combining Chained Bucket Hashing and Linear Hashing is the most efficient for streamed alerts with and without memory constraint, and (4) sort correlation with heap sort algorithm is the most efficient for alert correlation in batch.

Multivariate data visualization techniques are often limited in terms of the number of data records that can be simultaneously displayed in a manner that allows ready interpretation. Due to the size of the screen and number of pixels available, visualizing more than a few thousand data points generally leads to clutter and occlusion. This in turn restricts our ability to detect, classify, and measure phenomena of interest, such as clusters, anomalies, trends, and patterns. In this paper we describe our experiences in the development of multi-resolution visualization techniques for large multivariate data sets. By hierarchically clustering the data and displaying aggregation information for each cluster, we can examine the data set at multiple levels of abstraction. In addition, by providing powerful navigation and filtering operations, we can create an environment suitable for interactive exploration without overloading the user with dense information displays. In this paper, we illustrate that our hierarchical displays are general by successfully applying them to four popular yet non-scalable visualizations, namely parallel coordinates, glyphs, scatterplot matrices and dimensional stacking.

This paper describes new approaches to classification/identification of biological data. It is expected that the work: may be extensible to other domains such as the medical domain or fault diagnostic problems. Organisms are often classified according to the value of tests which are used for measuring some characteristic of the organism. When selecting a suitable test set it is important to choose one of minimum cost. Equally, when classification models are constructed for the posterior identification of unnamed individuals it is important to produce optimal models in terms of identification performance and cost. In this paper, we first describe the problem of selecting an economic test set for classification. We develop a criterion for differentiation of organisms which may encompass fuzzy differentiability. Then, we describe the problem of using batches of tests sequentially for identification of unknown organisms, and we explore the problem of constructing the best sequence of batches of tests in terms of cost and identification performance. We discuss how metaheuristic algorithms may be used in the solution of these problems. We also present an application of the above to the problem of yeast classification and identification.

In today’s software industry, design of black-box test cases is a manual activity, based mostly on human expertise, while automation tools are dedicated to execution of pre-planned tests only. However, the manual process of selecting test cases can rarely be considered as satisfactory both in terms of associated costs and the quality of produced software. This paper presents an attempt to automate a common task in black-box testing, namely reducing the number of combinatorial tests. Our approach is based on automated identification of relationships between inputs and outputs of a data-driven application. The set of input variables relevant to each output is extracted from execution data by a novel data mining algorithm called the info-fuzzy network (IFN). The proposed method does not require the knowledge of either the tested code, or the system specification, except for the list of software inputs and outputs. In the paper, we report the results of applying the proposed approach to a typical business application program.

Association mining algorithms can be used to discover sensitive knowledge from non-protected data that are voluntarily released for mining purposes. In this paper, we present a risk assessment procedure designed to control the security and privacy threats presented by association mining algorithms. The assessment procedure is based upon a new data analysis process that we introduce called Knowledge Hiding in Databases (KHD). The specific contributions of this paper include a description of the KHD process, and the development and implementation of a risk assessment procedure for association mining algorithms.

This paper presents a case study concerning scheduling and resource allocation issues in a spa. The paper is data-mining oriented. It discusses and describes how the history data can be used as a source for data-mining leading to discovery of rules or algorithms useful for prediction of resources requirements. In particular, we focused to identify groups of patients which appear frequently in the training set and which exhibit characteristic behavior or requirements of spa utilities. Then we predicted a set of health procedures to be passed for each member of such group. This approach resulted in a health procedure prediction algorithm satisfactory for early and convenient scheduling.

In this paper we present two methods for automated interoperability tests generation. The tests are applied to different entities implementing the same protocol and are peer-to-peer tests. They can be executed on test architectures either including only Points of Control and Observation (PCOs), or including both PCOs and Point of Observation (POs) to facilitate the diagnosis after the test execution. Both methods are applied in a case study to the interoperability test of the TCP/IP protocol. The protocol used for the experiment is TCP-Reno, which is specified in SDL. It is one of the most commonly referenced implementations. In the paper we present the tests sequences obtained and compare the experimental results of both methods.

Agent based intrusion detection systems (IDS) have advantages such as scalability, reconfigurability, and survivability. In this paper, we introduce a mobile-agent based IDS, called ABIDE (Agent Based Intrusion Detection Environment). ABIDE is comprised of various types of agents, all of which are mobile, lightweight, and specialized. The most common form of agent is the DMA (Data Mining Agent), which randomly moves around the network and collects information. The DMA then relays the information it has gathered to a DFA (Data Fusion Agent) which assesses the likelihood of intrusion. As we show in this paper, there is a quantifiable relationship between the number of DMA and the probability of detecting an intrusion. We study this relationship and its implications.

In this paper we present an algorithm for mining long patterns in databases. The algorithm finds large itemsets by using depth first search on a lexicographic tree of itemsets. The focus of this paper is to develop CPU-efficient algorithms for finding frequent itemsets in the cases when the database contains patterns which are very wide. We refer to this algorithm as DepthProject, and it achieves upto two orders of magnitude speedup over the recently proposed MaxMiner algorithm for finding long patterns. These techniques may be quite useful for applications in areas such as computational biology in which the number of records is relatively small, but the itemsets are very long. This necessitates the discovery of patterns using algorithms which are especially tailored to the nature of such domains.

Existing classification and rule learning algorithms in machine learning mainly use heuristic/greedy search to find a subset of regularities (e.g., a decision tree or a set of rules) in data for classification. In the past few years, extensive research was done in the database community on learning rules using exhaustive search under the name of association rule mining. The objective there is to find all rules in data that satisfy the user-specified minimum support and minimum confidence. Although the whole set of rules may not be used directly for accurate classification, effective and efficient classifiers have been built using the rules. This paper aims to improve such an exhaustive search based classification system CBA (Classification Based on Associations). The main strength of this system is that it is able to use the most accurate rules for classification. However, it also has weaknesses. This paper proposes two new techniques to deal with these weaknesses. This results in remarkably accurate classifiers. Experiments on a set of 34 benchmark datasets show that on average the new techniques reduce the error of CBA by 17% and is superior to CBA on 26 of the 34 datasets. They reduce the error of the decision tree classifier C4.5 by 19%, and improve performance on 29 datasets. Similar good results are also achieved against the existing classification systems, RIPPER, LB and a Naive-Bayes classifier.

In this paper we develop new techniques for predicting failures and monitoring in categorical event sequences. New techniques are needed because failures are rare and previous data mining algorithms were overwhelmed by the staggering number of very frequent, but entirely unpredictive patterns that exist in such databases. This paper combines several techniques for pruning out unpredictive and redundant patterns, which reduce the size of the returned rule set by more than three orders of magnitude. As a concrete application, we present PlanMine, an algorithm to extract patterns of events that predict failures in databases of plan executions. PlanMine has also been fully integrated into two real-world planning systems. We experimentally evaluate the rules discovered by PlanMine, and show that they are extremely useful for understanding and improving plans, as well as for building monitors that raise alarms before failures happen.

In this paper we describe how we can refine both objects and operations in an Object-Z specification. In particular, we will be concerned with changes of granularity of both objects and operations. Objects in that we wish to change the structure of objects in a specification. Operations in that we wish to provide explicit support for action refinement in this language. There are clear advantages in being able to change such levels of granularity when performing a refinement. In this paper we discuss the issues surrounding such refinements and derive general rules to support their use. We illustrate our ideas by looking at a specification of a cash point machine at a bank.

The paper addresses the problem of how to classify and diagnose the state of health of a river from the composition of its biological community. It is claimed that experts use two complex mental processes when interpreting such data, knowledge-based reasoning and pattern recognition. It is argued that existing classification methods are inadequate and that the application of advanced computer-based techniques is vital to the realisation of the full potential of biological monitoring. The paper then concentrates on a pattern recognition approach and demonstrates how Self Organising Maps (SOM), a type of unsupervised-learning neural network, can be used to classify and diagnose river quality. A brief introduction is given to the theory of SOMs and the interpretation of their output, as expressed in feature maps and class templates. SOMs are developed using two different methods of accounting for the confounding effects of environmental factors, and their relative performances are compared. Some improvements to the SOM architecture and functionality that are currently being implemented are briefly described, together with plans to use information theory for the assessment of performance. Finally, it is concluded that the methods of classification/diagnosis described in the paper have considerable potential not only in river quality monitoring, but also in other environmental fields.

Inference is a way to subvert access control mechanisms of database systems. Most existing work on inference detection relies on analyzing functional dependencies in the database schema. This paper is an extension to our earlier effort in developing a data level inference detection system [13]. In this paper, we introduce the split query inference rule, make an extension to the overlapping inference rule, and provide an in depth discussion on the applications of the inference rules on union queries. Data level inference detection is inevitably expensive. We have developed a prototype of the inference detection system to evaluate its performance. The result shows that the system performs better with larger number of attributes and records in the database, and smaller number of projected attributes and return tuples of the queries. Therefore, the inference detection system could be practical when users retrieve a small amount of data compare to the size of the database.

This paper is a first work on the link between two fields: validation and refinement in knowledge base systems and revision in knowledge representation. The validation evaluates knowledge base quality; among the criterion which can define this quality an interesting one is coherency. The coherency of a set of rules can be viewed as the potential consistency of this set of rules with respect to every possible input. Refinement is a way to modify an incoherent knowledge base in order to restore its coherency. A revision operation is an inference relation, which is defined in order to reason under inconsistency, it provides a way to select rules to delete in order to restore the consistency. In the two fields, usually the initial knowledge base is consistent and the inconsistency can only come from some input on which a deduction must be done. This paper underlines the link between refinement and revision by comparing consistency to coherency and then by defining an inference relation for refinement.

The use of OWA operators has been widely used in database contexts to process flexible queries where linguistic quantifiers are involved. This paper is devoted to comparing these operators’ use with other ways of addressing the same problems such as the Zadeh’s approach. Furthermore a new proposal is presented in order to avoid some inconveniences which the former approaches have in a wider class of problems. Finally, the paper analyzes the different types of queries with linguistic quantifiers which may appear in connection with relational databases and proposes the most appropiate way to solving them,in each case.

This paper provides a progress report on the development of methodology to estimate Nondestructive Evaluation (NDE) capability. The methodology uses combinations of physical modeling of an inspection process, along with laboratory and production data, to estimate Nondestructive Evaluation (NDE) capability. The methodology is based on a physical/statistical prediction model and will be used to predict Probability of Detection (POD), Probability of False Alarm (PFA) and Receiver Operating Characteristic (ROC) function curves. These output functions are used to quantify the NDE capability. The physical model will explain and allow predictions for the effects of making changes to the inspection setup (e.g. probe properties and scan increment). The statistical/empirical model will quantify unexplained variability, adjust for model bias, and provide a means for obtaining corresponding uncertainty intervals. Previous work on this project was reported in Meeker et al. (1996). The particular focus of this work is on the use of ultrasonic methods for detecting hard-alpha and other subsurface flaws in titanium using gated peak detection. This is a uniquely challenging problem because the inspection must detect very complex subsurface flaws in the presence of significant “material” noise. The underlying framework of the methodology should, however, be general enough to apply to other NDE methods. This paper describes recent work based on application of the new methodology to the detection of synthetic hard alpha flaws in titanium alloys.

Ultrasonic weld inspections are typically performed manually, which require significant operator expertise and time. Thus, automation of ultrasonic data analysis is an important area of current research in NDE. There is a need for automated data analysis schemes capable of handling imprecise data and providing results in real time. This paper presents a combination of neural networks and fuzzy-logic to automate different aspects of ultrasonic data analysis. Neural networks automate learning, and hence are best used when the relationship between the input space and the output space is highly nonlinear or unknown. The relationship between ultrasonic A-scan signal characteristics and defect class producing the signal is not straight forward. In this work a multi-layer perceptron is used for defect classification. Results form different feature extraction schemes icluding an unique combination of time- and frequency-domain features is presented. Fuzzy-logic automates knowledge representation using a fuzzy rule base. Hence, fuzzy-logic is best applied in situations where a knowledge base exists in the form of IF-THEN rules In this paper fuzzy-logic is applied to accept/reject criteria for weld evaluation. The advantages of using fuzzy-logic over traditional Boolean tree-based algorithms, for this application, are discussed.

This paper presents applications of graph theory to the design of graph matching neural networks for automated fingerprint identification. Given a sparse set of minutiae from a fingerprint image, complete with locations in the plane and (optionally) other labels such as ridge angles, ridge counts to nearby minutiae and so on, this approach to matching begins by constructing a graph-like representation of the minutiae map, utilizing proximity graphs, such as the sphere-of-influence graphs. These graph representations are more robust to noise such as translations, rotations and deformations. This paper presents the role of these graph representations in the design of graph matching neural networks for the matching and classification of fingerprint images.