Search

This paper presents an investigation into the use of the delay coordinate embedding technique with multi-input multi-output (MIMO) adaptive-network-based-fuzzy-inference system (ANFIS) to learn and predict the continuation of chaotic signals ahead in time. Based on the average mutual information and global false nearest neighbors techniques, the optimal values of the embedding dimension and the time delay are selected to construct the trajectory on the phase space. The MANFIS technique is trained by gradient descent algorithm. First, the parameter set of the membership functions is generated with the embedded phase space vectors using the back-propagation algorithm. Second, fine-tuned membership functions that make the prediction error as small as possible are built. The model is tested with both periodic and the Mackey-Glass chaotic time series. Moving root-mean-square error is used to monitor the error along the prediction horizon. [ABSTRACT FROM AUTHOR]

Determining the causal structure of a domain is frequently a key task in the area of Data Mining and Knowledge Discovery. This paper introduces ensemble learning into linear causal model discovery, then examines several algorithms based on different ensemble strategies including Bagging, Adaboost and GASEN. Experimental results show that (1) Ensemble discovery algorithm can achieve an improved result compared with individual causal discovery algorithm in terms of accuracy; (2) Among all examined ensemble discovery algorithms, BWV algorithm which uses a simple Bagging strategy works excellently compared to other more sophisticated ensemble strategies; (3) Ensemble method can also improve the stability of parameter estimation. In addition, Ensemble discovery algorithm is amenable to parallel and distributed processing, which is important for data mining in large data sets. [ABSTRACT FROM AUTHOR]

For many web usage mining applications, it is crucial to compare navigation paths of different users. This paper presents a reinforcement learning based method for mining the sequential usage patterns of user behaviors. In detail, the temporal data set about every user is constructed from the web log file, and then the navigation paths of the users are modelled using the extended Markov decision process. The proposed method could learn the dynamical sequential usage patterns on-line. [ABSTRACT FROM AUTHOR]

Fundamental astronomical questions on the composition of the universe, the abundance of Earth-like planets, and the cause of the brightest explosions in the universe are being attacked by robotic telescopes costing billions of dollars and returning vast pipelines of data. The success of these programs depends on the accuracy of automated real time processing of the astronomical images. In this paper the needs of modern astronomical pipelines are discussed in the light of fuzzy-logic based decision-making. Several specific fuzzy-logic algorithms have been develop for the first time for astronomical purposes, and tested with excellent results on data from the existing Night Sky Live sky survey. [ABSTRACT FROM AUTHOR]

The on-line update of classifiers is an important concern for categorizing the time-varying neurophysiological signals used in brain computer interfaces, e.g. classification of electroencephalographic (EEG) signals. However, up to the present there is not much work dealing with this issue. In this paper, we propose to use the idea of gradient decorrelation to develop the existent basic Least Mean Square (LMS) algorithm for the on-line learning of Bayesian classifiers employed in brain computer interfaces. Under the framework of Gaussian mixture model, we give the detailed representation of Decorrelated Least Mean Square (DLMS) algorithm for updating Bayesian classifiers. Experimental results of off-line analysis for classification of real EEG signals show the superiority of the on-line Bayesian classifier using DLMS algorithm to that using LMS algorithm. [ABSTRACT FROM AUTHOR]

Grapheme-to-phoneme (G2P) conversion is an important subcomponent in many speech processing systems. The difficulty in Chinese G2P conversion is to pick out one correct pronunciation from several candidates according to the context information such as part-of-speech, lexical words, length of the word, or position of the polyphone in a word or a sentence. By evaluating the distribution of polyphones in a large text corpus with correct pinyin transcriptions, this paper points out that correct G2P conversion for 78 key polyphones greatly decrease the overall error rate. This paper proposed a fast Transformation-based error-driven learning (TBL) algorithm to solve G2P conversion. The correct rates of polyphones, which originally have high accuracy or low accuracy, are both improved. After compared with Decision Tree algorithm, TBL algorithm shows better performance to solve the polyphone problem. [ABSTRACT FROM AUTHOR]

Acquisition of the correspondences from the image sequences with high-efficiency and high-precision is a fundamental and key problem in computer virtual techniques and human-computer interaction systems. In this paper, we start from image topological structures, aiming firstly finding the correspondences between topological structures of the image sequences. Then, we further attain accurate correspondences between feature points by adopting local search methods. In order to speed up the search process for desired data, a grid technique is introduced and some new concepts, such as SQVSBS, and related theories are put forward. The specific characteristics of the algorithm are: (1) using the Top-to-Bottom strategy, from rough estimates to accurateness, from local to global; (2) getting better time complexity, O(Max(f2,A)), which is better than that that given in the references [11] and [12],here, f is the number of feature grids; (3) avoiding wrong matches deriving from local optical solutions; and (4) Focusing on the internal topological relationships between feature griddings upon which the characteristic of a feature gridding is based. [ABSTRACT FROM AUTHOR]

This paper introduces a novel transductive knowledge based fuzzy inference system (TKBFIS) and its application for creating personalized models. In transductive systems a local model is developed for every new input vector, based on some closest data to this vector from the training data set. A higher-order TSK type fuzzy inference engine is applied in TKBFIS. Some existing formulas or equations, which are used to represent the knowledge and usually have a non-linear form, are taken as consequent parts of the fuzzy rules. The TKBFIS uses a gradient descent algorithm for its training. In this paper, the TKBFIS is illustrated with a case study of personalized modeling for renal function estimation of patients and the result is compared with other transductive or inductive methods. [ABSTRACT FROM AUTHOR]

Clustering results often critically depend on density and similarity, and its complexity often changes along with the augment of sample dimensionality. In this paper, we refer to classical shared nearest neighbor clustering algorithm (SNN), and provide a high-dimensional shared nearest neighbor clustering algorithm (DSNN). This DSNN is evaluated using a freeway traffic data set, and experiment results show that DSNN settles many disadvantages in SNN algorithm, such as outliers, statistic, core points, computation complexity etc, also attains better clustering results on multi-dimensional data set than SNN algorithm. [ABSTRACT FROM AUTHOR]

In this paper, a support vector machine (SVM) multi-user receiver based on competition learning (CL) strategy is proposed. The new algorithm adopts a heuristic approach to extend standard SVM algorithm for multiuser classification problem, and also a clustering analysis is applied to reduce the total amount of computation. In implementation of multi-user receiver, an asymptotical iterative algorithm is used to guide the learning of the input sample pattern. The digital result shows that the new multi-user detector scheme has a relatively good performance comparing with the conventional MMSE detector especially under the heavy interference environment. [ABSTRACT FROM AUTHOR]

The major problems of currently used approaches for extracting symbolic rules from trained neural networks are analyzed. The lack of efficient heuristic information is the fundamental reason that causes the low effectiveness of currently used approaches. In this paper, a new measurement method of the classification power of attributes on the basis of differential information of the trained neural networks is proposed, which is suitable for both continuous attributes and discrete attributes. Based on this new measurement method, a new approach for rule extraction from trained neural networks and classification problems with continuous attributes is proposed. The performance of the new approach is demonstrated by several computing cases. The experimental results prove that the approach proposed can improve the validity of the extracted rules remarkably compared with other rule extracting approaches, especially for the complicated classification problems. [ABSTRACT FROM AUTHOR]

This paper introduces some widely used techniques related to nude image detection. By analyzing the merits and drawbacks of these techniques, a new nude image detection method is proposed. The proposed approach consists of two parts: the content-based approach, which aims to detect the nude image by analyzing whether it contains large mass of skin region, and the image-based approach, which extracts the color and spatial information of the image using the color histogram vector and color coherence vector, and makes classification based on the CHV and CCVs of the training samples. From the experimental results, our algorithm can achieve a classification accuracy of 85% with less than 10% false detection rate. [ABSTRACT FROM AUTHOR]

Bayesian networks and undirected graphical models are often used to cope with uncertainty for complex systems with a large number of variables. They can be applied to discover causal relationships and associations between variables. In this paper, we present heuristic algorithms for structural learning of undirected graphical models from observed data. These algorithms are applied to traditional Chinese medicine. [ABSTRACT FROM AUTHOR]

HTTP request exploitations take substantial portion of network-based attacks. This paper presents a novel anomaly detection framework, which uses data mining technologies to build four independent detection models. In the training phase, these models mine specialty of every web program using web server log files as data source, and in the detection phase, each model takes the HTTP requests upon detection as input and calculates at least one anomalous probability as output. All the four models totally generate eight anomalous probabilities, which are weighted and summed up to produce a final probability, and this probability is used to decide whether the request is malicious or not. Experiments prove that our detection framework achieves close to perfect detection rate under very few false positives. [ABSTRACT FROM AUTHOR]

This paper proposes a hybrid system which combines computer extracted features and human interpreted features from the mammogram, with the statistical classifier's output as another kind of features in conjunction with a genetic neural network classifier. The hybrid system produced better results than the single statistical classifier and neural network. The highest classification rate reached 91.3%. The area value under the ROC curve is 0.962. The results indicated that the mixed features contribute greatly for the classification of mass patterns into benign and malignant. [ABSTRACT FROM AUTHOR]

As one kind of image having strong texture character, ridge distance is the important attribute of fingerprint image. It is important to estimate the ridge distance correctly for improving the performance of the automatic fingerprint identification system. The traditional Fourier transform spectral analysis method had the worse redundancy degree in estimating the ridge distance because it was based on the two-dimension discrete Fourier spectrum. The paper introduces the sampling theorem into the fingerprint image ridge distance estimation method, transforms the discrete spectrum into two-dimension continuous spectrum and obtains the ridge distance on the frequency field. The experimental results indicate that the ridge distance obtained from this method is more accurate and has improved the rate of accuracy of the automatic fingerprint identification system to a certain extent. [ABSTRACT FROM AUTHOR]

The attribute oriented induction (AOI) is a kind of aggregation method. By generalizing the attributes of the alert, it creates several clusters that includes a set of alerts having similar or the same cause. However, if the attributes are excessively abstracted, the administrator does not identify the root cause of the attack. In addition, deciding time interval of clustering and deciding min_size are one of the most critical problems. In this paper, we describe about the over-generalization problem because of the unbalanced generalization hierarchy and discuss the solution of the problem. We also discuss problem to decide time interval and meaningful min_size, and propose reasonable method to solve these problems. [ABSTRACT FROM AUTHOR]

In this paper, we present a novel method to find the right expert who matches a certain project well. The idea behind this method includes building domain ontologies to describe projects and experts and calculating similarities between projects and domain experts for matching. The developed system consists of four main components: ontology building, document formalization, similarity calculation and user interface. First, we utilize Protégé to develop the predetermined domain ontologies in which some related concepts are defined. Then, documents concerning experts and projects are formalized by means of concept trees with weights. This process can be done either automatically or manually. Finally, a new method that integrates node-based and edge-based approach is proposed to measure the semantic similarities between projects and experts with the help of the domain ontologies. The experimental results show that the developed information matching system can reach the satisfied recall and precision. [ABSTRACT FROM AUTHOR]

Existing sketch recognition algorithms are mainly on recognizing single segments or simple geometric objects (such as rectangles) in a stroke. We present in this paper an on-line sketch recognition algorithm for composite shapes. It can recognize single shape segments such as straight line, polygon, circle, circular arc, ellipse, elliptical arc, hyperbola, and parabola curves in a stroke, as well as any composition of these segments in a stroke. Our algorithm first segments the stroke into multi-segments based on a key point detection algorithm. Then we use "combination" fitting method to fit segments in sequence iteratively. The algorithm is already incorporated into a hand sketching based modeling prototype, and experiments show that our algorithm is efficient and well suited for real time on-line applications. [ABSTRACT FROM AUTHOR]

Web prefetching technique is one of the primary solutions used to reduce Web access latency and improve the quality of service. This paper makes use of Zipf's 1st law and Zipf's 2nd law to model the Web objects' popularity, where Zipf's 1st law is employed to model the high frequency Web objects and 2nd law for the low frequency Web objects, and proposes a PPM prediction model based on Web objects' popularity for Web prefetching. A performance evaluation of the model is presented using real server logs. Trace-driven simulation results show that not only the model is easily to be implemented, but also can achieve a high prediction precision at the cost of relative low storage complexity and network traffic. [ABSTRACT FROM AUTHOR]

In the artificial neural networks (ANNs), feature selection is a well-researched problem, which can improve the network performance and speed up the training of the network. The statistical-based methods and the artificial intelligence-based methods have been widely used to feature selection, and the latter are more attractive. In this paper, using genetic algorithm (GA) combining with mutual information (MI) to evolve a nearoptimal input feature subset for ANNs is proposed, in which mutual information between each input and each output of the data set is employed in mutation in evolutionary process to purposefully guide search direction based on some criterions. By examining the forecasting at the Australian Bureau of Meteorology, the simulation of three different methods of feature selection shows that the proposed method can reduce the dimensionality of inputs, speed up the training of the network and get better performance. [ABSTRACT FROM AUTHOR]

Supervised local tangent space alignment is proposed for data classification in this paper. It is an extension of local tangent space alignment, for short, LTSA, from unsupervised to supervised learning. Supervised LTSA is a supervised dimension reduction method. It make use of the class membership of each data to be trained in the case of multiple classes, to improve the quality of classification. Furthermore we present how to determine the related parameters for classification and apply this method to a number of artificial and realistic data. Experimental results show that supervised LTSA is superior for classification to other popular methods of dimension reduction when combined with simple classifiers such as the k-nearest neighbor classifier. [ABSTRACT FROM AUTHOR]

Much evidence from visual psychology suggests that images can be looked as a kind of language, by which image semantics can be unambiguously expressed. In this paper, we discuss the primitives and grammar of image language based on cognition theory. Hence image understanding can surely be manipulated in the same way as language analysis. Keywords: Image semantics, Visual cognition theory, Image Language. [ABSTRACT FROM AUTHOR]

Human vision system can understand images quickly and accurately, but it is impossible to design a generic computer vision system to challenge this task at present. The most important reason is that computer vision community is lack of effective collaborations with visual psychologists, because current object recognition systems use only a small subset of visual cognition theory. We argue that it is possible to put forward a generic solution for image object recognition if the whole vision cognition theory of different schools and different levels can be systematically integrated into an inherent computing framework from the perspective of computer science. In this paper, we construct a generic object recognition solution, which absorbs the pith of main schools of vision cognition theory. Some examples illustrate the feasibility and validity of this solution. Keywords: Object recognition, Generic solution, Visual cognition theory, Knowledge. [ABSTRACT FROM AUTHOR]

In this paper we propose using manifolds modeled by probabilistic principle surfaces (PPS) to characterize and classify high-D data. The PPS can be thought of as a nonlinear probabilistic generalization of principal components, as it is designed to pass through the "middle" of the data. In fact, the PPS can map a manifold of any simple topology (as long as it can be described by a set of ordered vector co-ordinates) to data in high-dimensional space. In classification problems, each class of data is represented by a PPS manifold of varying complexity. Experiments using various PPS topologies from a 1-D line to 3-D spherical shell were conducted on two toy classification datasets and three UCI Machine Learning datasets. Classification results comparing the PPS to Gaussian Mixture Models and K-nearest neighbours show the PPS classifier to be promising, especially for high-D data. [ABSTRACT FROM AUTHOR]

This paper describes the development of an Automated Knowledge Extraction Agent (AKEA) which was designed to acquire online news and document from the internet for the establishment of a knowledge based crisis communication portal. It was recognized that in times of crisis, an effective communication mechanism is essential to maintain peace and calmness in the community by providing timely and appropriate information. It is proposed that the incorporation of software agents into the crisis communication portal will be capable to send alert news to subscribed users via internet and mobile services. The proposed system consists of crawler, wrapper, name-entity tagger, AIML (Artificial Intelligence Markup language) and an animated character is used in the front-end for human computer communication. [ABSTRACT FROM AUTHOR]

The Web can be regarded as a huge graph when each Web page is regarded as a node and each hyperlink as an edge. There are several attempts for visualizing the structure of the Web, such as touchgraph or KartOO. In order to achieve visualization that assists users' information acquisition from the Web, two constructs (keywords and pages) are required in the visualization. In this paper, a cluster of keywords and Web pages is regarded as "structural information" in the Web. We have developed a visualization system that shows clusters of Web pages and keywords. Based on online Web resources, appropriate relations can be visualized without analyzing the contents of Web pages. [ABSTRACT FROM AUTHOR]

With the dramatic development of web technologies, tremendous amount of information become available to users. The great advantages of the web are the ease with which information can be published and made available to a wide audience, and the ability to organize and connect different resources in a graph-based structure using hyperlinks. However, most of these links are created manually and the page that the link represents must be known to the author of the link. In this paper, we propose a decision-tree-based approach to solve this problem. We set up a system that gathers information about the candidate pages, evaluates them and creates links to them automatically. [ABSTRACT FROM AUTHOR]

Classifiers, as one of the important tools of analyzing gene expression data in the post-genomic epoch, have been used widely in the classification of different cancer types in the past few years. Although most existing classifiers have high classification accuracy, the process of classification is a black box and they can not give biologists more information and interpretable results of classification. In this paper, we propose a novel visualization cancer classification method. Besides offering high classification accuracy, the method can help us identify complex disease-related genes and assess gene expression variation during the process of classification. The results of classification are natural and interpretable and the process of classification is visible. To evaluate the performance of the method we have applied the proposed method to three public data sets. The experimental results demonstrate that the approach is feasible and useful. [ABSTRACT FROM AUTHOR]

This paper addresses a weighted dynamic localized clustering unique to a hierarchical sensor network structure, while reducing the energy consumption of cluster heads and as a result prolonging the network lifetime. Low-Energy Localized Clustering, our previous work, dynamically regulates the radii of clusters to minimize energy consumption of cluster heads while the network field is being covered. We present weighted Low-Energy Localized Clustering (w-LLC), which consumes less energy than LLC with weight functions. [ABSTRACT FROM AUTHOR]

Core operations (e.g. TDMA scheduler, synchronized sleep period, data aggregation) of many proposed protocols for different layer of sensor network necessitate clock synchronization. Our paper mingles the scheme of dynamic clustering and diffusion based asynchronous averaging algorithm for clock synchronization in sensor network. Our proposed algorithm takes the advantage of dynamic clustering and then applies asynchronous averaging algorithm for synchronization to reduce number of rounds and operations required for converging time which in turn save energy significantly than energy required in diffusion based asynchronous averaging algorithm. [ABSTRACT FROM AUTHOR]

This paper proposed a knowledge-based fault diagnosis system using ORP (Oxidation-Reduction Potential) and DO (Dissolved Oxygen) values which usually applied as control parameters in wastewater treatment plants. If the basic control parameters such as ORP and DO can be applied to operation diagnosis, the stability of process will be remarkably improved without additional expenses. This proposed diagnosis method uses only the ORP and DO values obtained from full-scale SBR (Sequencing Batch Reactor). For the classification and diagnosis of these statues, a sequenced process of preprocessing, dimension reducing using PCA and feature extraction with ORP, DO and a synthetic parameter of [ORP DO] were proposed and applied. As results, the synthetic parameter of [ORP DO] shows better fault recognition rate than that of independent application of each parameter. It was considered that this diagnostic system using control parameters could be used to support small-scale wastewater treatment management. [ABSTRACT FROM AUTHOR]

The Systems Biology Markup Language(SBML) is an XML-based format for representing models of Systems Biology. BioSPI is a formal model to simulate biological systems, which is evolved from process calculi. Based on the previous research on modeling Systems Biology using process algebra, we propose a method to map SBML to BioSPI. The motivation of the work is to make full use of BioSPI to analyze biological systems described by SBML. In this paper, the mapping rules are presented and an example is given to show the simulation results. [ABSTRACT FROM AUTHOR]

This paper proposes a method of detecting shot transitions by compensating for motions contained in video sequences. The proposed method detects shot transitions including cuts, fades, and dissolves after compensating for global motions (camera motions) and eliminating local motions (moving objects), so that our approach prevents false positives caused by those motions. Experimental results show that our method works as a promising solution. [ABSTRACT FROM AUTHOR]

Wireless sensor networks will become very useful in the near future. The efficient energy consumption in wireless sensor network is a critical issue since the energy in the nodes is constrained resource. In this paper, we present a transmission relay method of communications between BS (Base Station) and CHs (Cluster Heads) for balancing the energy consumption and extending the average lifetime of sensor nodes by the fuzzy logic application. The proposed method is designed based on LEACH protocol. The area deployed by sensor nodes is divided into two groups based on distance from BS to the nodes. RCH (Relay Cluster Head) relays transmissions from CH to BS if the CH is in the area far away from BS in order to reduce the energy consumption. RCH decides whether to relay the transmissions based on the threshold distance value that is obtained as a output of fuzzy logic system. Our simulation result shows that the application of fuzzy logic provides the better balancing of energy depletion and prolonged lifetime of the nodes. [ABSTRACT FROM AUTHOR]

To study the spatial relationships with the instability is becoming one of the hot spots and the difficulties in studying the spatial relationships. This paper express and apply the information of relationships among spatial objects in the real world in computer system from the cognitive view, study the fuzzy extension about description of spatial relationships at the base. Guided by the spatial query, we makes the model on the base of regular indefinite spatial inferring, applies fuzzy theory and spatial relationship theory in the spatial query and solves the fuzzy location problems in applying GIS network resource management. [ABSTRACT FROM AUTHOR]

In a production process, there are numerous systems that provide information/reports for various purposes. However, most of the knowledge for decision-making is kept in minds of experienced employees rather than exists in IT systems that can be managed systematically. Even experienced managers may make flaw/improper decisions due to the lack of must-known information, not to mention what those who are less experienced or have been urged by the pressure of time will probably do. In this paper, a fuzzy-logic-based functional center hierarchical model named Dynamic Master Logic (DML) is designed as an interview interface for representing engineers' tacit knowledge and a self-learning model for tuning the knowledge base from historical cases. The DML representation itself can also be the inference engine in a manufacture process diagnoses expert system. A semiconductor Wafer Acceptance Test (WAT) root cause diagnostics which usually involves more than 40,000 parameters in a 500-step production process is selected to examine the DML model. In this research, it has been proven to shorten the WAT diagnostics time from 72 hours to 15 minutes with 98.5% accuracy and to save the human resource form 2 senior engineers to one junior engineer. [ABSTRACT FROM AUTHOR]

This paper presents the feature recognition technique that recognizes the features from 2D ship drawings using the fuzzy inference system. Generally, ship drawings include a lot of symbols and texts. They were complicatedly combined each other. So, it is very difficult to recognize the feature from 2D ship model. The fuzzy inference system is suitable to solve these problems. Input information for fuzzy inference is connection type of drawing elements and properties of element. Output value is the correspondence between target feature and candidate feature. The recognition rule is the fuzzy rule that has been predefined by designer. In this study, the midship section drawing of general cargo ship was used to verifying suggested methodology. Experimental results showed that this approach is more efficient than existing methods and reflects the human knowledge for recognition of the feature. [ABSTRACT FROM AUTHOR]

Reliable localization is a fundamental issue in robot navigation techniques. This paper describes an apporach for realizing self-localization of mobile robot by matching the local map generated from a 2D laser scanner. Environment map is represented by occupancy grids and it fuses the information of the robot's pose using dead-reckoning method and the range to obstacles by laser scanner using maximum likehood estimation. After a current laser scan, the positon of mobile robot, in relation to a previous scan and pose estimates, is computed by matching the local map using fuzzy logic method. The effectiveness of this method is demonstrated by experiments. [ABSTRACT FROM AUTHOR]

Product development process can be viewed as a set of subprocesses with stronger interrelated dependency relationships. In this paper, the quantitative and qualitative dependency measures of serial and parallel product development processes are analyzed firstly. Based on the analysis results, the process net is developed where the processes are viewed as nodes and the logic constraints are viewed as verges of the net. The fuzzy-logic-based reasoning mechanism is developed to reason the dependency relations between development processes in the case that there is no sufficient quantitative information or the information is fuzzy and imprecise. The results show that the proposed method can improve the reasoning efficiency, reduce the cost and complexity degree of process improvement, and make a fast response to the dynamic development environment. [ABSTRACT FROM AUTHOR]

This paper presents an evaluation of the four of the more common gene finding programs. The evaluation was conducted on a new data set consisting of only human genome sequences extracted from GenBank. Newest sequences were used to avoid overlap with the training sets of the gene-finding programs. The results of this evaluation are then used to classify the gene finding programs using fuzzy logic. The programs are classified into three fuzzy sets of high, mediocre and low accuracy. The results are then presented in the form of words so as to be easily understood by humans. [ABSTRACT FROM AUTHOR]

Fuzzy neural networks have been successfully applied to analyze/generate predictive rules for medical or diagnostic data. This paper presents selected membership functions extracted by a fuzzy neural network named NEWFM. The selected membership functions can capture the concentrated and essential information without sacrificing the classification capability. To verify the performance of the NEWFM, the well-known data set of Wisconsin breast cancer is performed. We applied NEWFM model to extract fuzzy membership functions for the UCI antibody deficiency syndrome diagnosis. Then selected features obtained by non-overlapped area measurement method are presented. [ABSTRACT FROM AUTHOR]

An automatic segmentation technique for microscopic bone marrow white blood cell images is proposed in this paper. The segmentation technique segments each cell image into three regions, i.e., nucleus, cytoplasm, and background. We evaluate the segmentation performance of the proposed technique by comparing its results with the cell images manually segmented by an expert. The probability of error in image segmentation is utilized as an evaluation measure in the comparison. From the experiments, we achieve good segmentation performances in the entire cell and nucleus segmentation. The six-class cell classification problem is also investigated by using the automatic segmented images. We extract four features from the segmented images including the cell area, the peak location of pattern spectrum, the first and second granulometric moments of nucleus. Even though the boundaries between cell classes are not well-defined and there are classification variations among experts, we achieve a promising classification performance using neural networks with five-fold cross validation. [ABSTRACT FROM AUTHOR]

Recently, there are remarkable progress in similarity computing for 3D geometric models. Few focus is put on the research of the similarity between volumetric models. This paper proposes a novel approach for performing similarity computation between two volumetric data sets. For each data set, it is performed by four stages. First, the volume data set is resampled into a unified resolution. Second, the data set is band-pass filtered and quantized to reveal its physical attributes. The resulting voxels are then normalized into a canonical coordinate system concerning the center of mass and scale. Subsequently, a series of uniformly spaced concentric shells around the center of mass are constructed, based on which spherical harmonics analysis (SHA) is applied. The coefficients of SHA constitute a rotation invariant spectrum descriptor which are used to measure the similarity between two data sets. The algorithm has been performed on a set of clinical CT and MRI data sets and the preliminary results are fairly inspiring. [ABSTRACT FROM AUTHOR]

In this paper, we apply data mining techniques to construct intrusion detection patterns. We mine both system audit data and network traffic data for consistent and useful patterns of program and user behavior, and use an iterative low-frequency-finder mining algorithm to find the low frequency but important patterns. [ABSTRACT FROM AUTHOR]

Video abstraction can be useful in multimedia database indexing and querying and can illustrate the important content of a longer video to quick browsing. Further, in sports video, replay scene often demonstrates the highlight of the video. The detection of replay scene in the sports video is a key clue to sports video summarizing. In this paper, we present a framework of replay scene based video abstraction in MPEG sports video. Moreover, we detect identical events using color and camera information after detecting replay scene using MPEG feature. At last, we propose a three-layer replay scene based sports video abstraction. It can achieve real time performance in the MPEG compressed domain, which is validated by experimental results. [ABSTRACT FROM AUTHOR]

Attribute uncertainty may support decision-making and measure the reliability of GIS analysis. This paper presents the attribute uncertainty in GIS data, with a complete perspective of concepts, sources, nature and applicable tools. And some valuable research contents are further given. [ABSTRACT FROM AUTHOR]

This paper presents a framework for automatically supplying keyphrases for a Chinese news document. It works as follows: extracts Chinese character strings from a source article as an initial set of keyphrase candidates based on frequency and length of the strings, then, filters out unimportant candidates from the initial set by using elimination-rules and transforms vague ones into their canonical forms according to controlled synonymous terms list and abbreviation list, and finally, selects the best items from the set of the remaining candidates by score measure. The approach is tested on People Daily corpus and the experiment results are satisfactory. [ABSTRACT FROM AUTHOR]

Intrusion detection is a critical component of secure information system. Recently applying artificial intelligence, machine learning and data mining techniques to intrusion detection system are increasing. But most of researches are focused on improving the classification performance of classifier. Selecting important features from input data lead to a simplification of the problem, faster and more accurate detection rates. Thus selecting important features is an important issue in intrusion detection. Another issue in intrusion detection is that most of the intrusion detection systems are performed by off-line and it is not proper method for realtime intrusion detection system. In this paper, we develop the realtime intrusion detection system which combining on-line feature extraction method with Least Squares Support Vector Machine classifier. Applying proposed system to KDD CUP 99 data, experimental results show that it have remarkable feature feature extraction and classification performance compared to existing off-line intrusion detection system. [ABSTRACT FROM AUTHOR]

This paper presents an analysis of product quality improvement in ultra-precision manufacturing industry using data mining for developing quality improvement strategies. Based on 11320 ultra-precision optical products that were produced from the study factory during the period of June 1 and August 31, 2004, important factors impacting the product quality were identified via the decision tree method for data mining. Findings showed that the important factors for the percentage of defectives were type of processing chain, precision requirement, product classes, and raw material. The optimum range of target group in production quality indicators was identified from the gains chart. [ABSTRACT FROM AUTHOR]