term subnetwork and the document subnetwork. The former’s nodes represent indexing terms. Nodes links are used to depict dependence relationships between indexing terms. The latter’s nodes represent the set of documents. The relationships between a document and its indexing terms are presented by the links between the two subnetworks. In this model there is no node for the user’s query. In fact, query terms are considered as evidence that should be introduced into the system. To reduce the computation cost, this model uses canonical models instead of learning algorithms to estimate the conditional probability distribution of nodes. Canonical models are also used to do inference process. In a similar approach, Dongyu et al  proposed a model having almost the same as the BNRM’s one. However, the set of arcs are not oriented in the term subnetwork. This constitutes the main difference between the two models. A third model, proposed by de Campos et al , uses two term-layers to encode term relationships. It is based on the use of a term clustering technique to extract the strongest relationships among terms. Therefore, the complete Bayesian network contains three simple layers: two term layers and a document layer. A fourth model with two terms, layers was proposed by Xu et al . Here, the term relationships are mined by using word similarity extracted from a thesaurus.
Abstract: This paper presents a new control methods based on adaptive Neuro-Fuzzy damping controller and adaptive Artificial Neural Networks damping controller techniques to control a Unified Power Flow controller (UPFC) installed in a single machine infinite bus Power System. The objective of Neuro-Fuzzy and ANN based UPFC controller is to damp power system oscillations.Phillips-Herffron model of a single machine power system equipped with a UPFC is used to model the system. In order to damp power system oscillations, adaptive neuro-fuzzy damping controller and adaptive ANN damping controller for UPFC are designed and simulated. Simulation is performed for various types of loads and for different disturbances. Simulation results demonstrate that the developed adaptive ANN damping controller has an excellent capability in damping electromechanical oscillations which exhibits a superior damping performance in comparison to the neuro-fuzzy damping controller as well as conventional lead-lag controller.
Wireline acquisitions in open holes provide petrophysical data for oil exploration purposes. In these operations, sev- eral physicochemical properties of rocks (i.e. rock + fluid) are recorded, such as resistivity, density, porosity, radioactivity, and others. Petrophysical evaluation both of electrical and image pro- files allows, for example, to calculate porosity and permeability of rocks, as well as water saturation. This information is essential in the petroleum exploratory chain, since it assists on defining target areas and guiding production of oil fields. Indirect litho- logic inference is another important product of well logs interpre- tation. Lithologies retrieved from petrophysical interpretation are known as lithofacies.
intelligent context aware vertical handover decision algorithm for multimode mobile terminals which takes the session transfer into account considering the intelligence both on terminal and network side which they term as context information. The algorithm is based on the Analytical Hierarchy Process (AHP) for which the inputs are defined using primary objectives for decision algorithm. The objectives defined in turn depend on the user preferences. The capabilities of the available networks are measures and checked for the suitability of the user preferences and the b e s t suitable n e t wo r k is selected. This method suits only for average us e r s , minimizing processing time, hand over delay and CPU and memory usage.
propose a new modelbased on high-order and fuzzy-trend of logical relationships for forecasting. In , the particle swarm optimization technique is exploited in the forecasting model to improve the forecasting accuracy. In , the fuzzy logical relationships are replaced with the artificial neural networks. In , a forecasting modelbased on similarity measures of fuzzy logical relationships is proposed. In , a method of partitioning the fuzzy logical relationships based on support vector machine is proposed. In , , , multiple variables time series are considered simultaneously to improve the forecasting accuracy, where multiple variables include a main factor and at least a secondary factor. In , an adaptive selection of analysis windows and heuristic rules is proposed to improve forecasting accuracy. In , a forecasting model of fuzzy time series which exploits respectively particle swarm optimization algorithm and fuzzy K-means clustering algorithm to obtain the optimum partition of the universe of discourse is proposed. The aforementioned works, in the testing phase, mainly exploit a certain order of the fuzzy relationship to forecast values, regardless of fixed order (first-order, second-order or higher-order) or adaptive order of fuzzy relationships is used for forecasting. However, different orders of the fuzzy relationships have different information, i.e., the lower orders fuzzy logical relationships have robustness information, while the higher orders fuzzy logical relationships have precision information. Only considering a single order of the fuzzy relationship in the testing phase will not make full use of effective information of different orders.
In order to test the classiﬁer, a software framework was implemented. The frame- work can be conﬁgured using XML ﬁle, where the user can deﬁne the components and references among them. Therefore, the components may be tuned without need of code recompilation. Based on XML ﬁle content, appropriate dataset is loaded and tests are conducted. The constructed system is extensible and other classiﬁers can be included – currently besides neuro-fuzzy classiﬁer, there are neural networks and k nearest neighbours classiﬁer. Two “ensemble classiﬁers” – SAMME boosting and bagging, were also implemented. The system was written in Java language. To im- plement environment, the Spring Framework 1 has been used. One of its well-known
When user wants to search for a web service for a specific purpose he should write his query in English language because the service metadata is written in English language. The motivation for addressing this idea came from our experience in developing the web service and studying the Cross- Language InformationRetrieval (CLIR) methods and data mining techniques. We deal with the web services as collections of documents that should be prepared for IR techniques, then we applied CLIR techniques to find the suitable service that matches the user query that written in other language. The process that is responsible for finding the suitable service is matchmaking process.it is the process of finding suitable services given by the providers for the service requests of consumers. The current service discovery mechanism of WSs is based on WSDL  and UDDI . WSDL is an XML based language to describe properties of services that written in English language. UDDI is a registry where service providers can advertise their services and service consumers can search for services. The specific objective of our research is therefore to apply CLIR in the web services discovery; this is done by modifying the Match Maker process by adding CLIR components to support the Cross language web service discovery.
BoF generative approaches take an intermediate step to classification. They first model the features probability of each class before proceeding to classification – and in the BoF case, assume that the low level descriptors that represent each song are independent and identically distributed (i.i.d). Many of the early works opted for this generative approach, for example in [Logan and Salomon, 2001] a k-means algorithm was used to model spectral similarity be- tween songs (and genres), and in [Aucouturier and Pachet, 2002] a similar strategy for genre classifications was done using Gaussian mixture models (GMMs) trained with the classical Ex- pectation Maximization (EM) algorithm [Dempster et al., 1977] (see Section 3.1 for details of the two methods). In [Logan and Salomon, 2001] to measure distance between models the authors use the earth movers distance [Rubner et al., 2000], a technique used to compare two cluster representations. Clusters are analogous to “piles of earth” and the amount of “earth” (probability mass) necessary to move in order to convert one set of clusters into another mea- sures the (dis)similarity between the two models. The GMMs are an estimate of the feature distributions, and an information theoretic way of measuring the differences between distribu- tions is via a symmetrized version of the Kullback-Leibler (KL) divergence (Equation 3.23). In practice some approximations are necessary for the GMM case, and in [Aucouturier and Pa- chet, 2002] Monte Carlo sampling is used for that purpose, although other strategies are also available [Vasconcelos, 2001]. GMMs have been extensively used in MIR for audio-based clas- sification and similarity estimation: for e.g. in [Aucouturier and Pachet, 2004, Berenzweig et al., 2004, Burred and Lerch, 2003, Flexer et al., 2005, Jensen et al., 2009, Jiang et al., 2002, Turnbull et al., 2008b, Tzanetakis and Cook, 2002]. With probabilistic models, one can also de- termine class memberships via maximum likelihood or maximum a posteriori estimation. One model is build for each class and new songs are attributed to the class with the highest likelihood (or maximum a posteriori probability). For classification, the models are built using data from songs from the same class, but it is also straightforward to estimate them based on the feature vectors from individual songs. Therefore, this type of approach can be used for more general music similarity measures (i.e. similarity between songs). For comparing two songs, one can either measure similarity via KL-divergence, or via maximum likelihood – i.e. compute the likelihood of the features from Song A using the model from Song B . Note that this is possible,
Marking content with descriptive terms, also called key- words or tags, is a common way of organizing content for future navigation, filtering or search. Collaborative tagging is the practice of allowing anyone, especially consumer, to freely attach keywords or tags to con- tent . Collaborative tagging systems have become valuable tools for sharing and exploring content, where tag-item associations can be aggregated over thousands or even millions of users. In  is designed and ac- complished a study of a personalized annotation man- agement system (PAMS 2.0) for managing, sharing, and reusing individual and collaborative annotations. Through PAMS 2.0 the authors investigate the effects of different annotation sharing scenarios on quantity of annotation and its influence on learning achievements. The results show that annotation process on learning achievements becomes effective and the sharing mech- anism is positive for the majority of students. In  the authors elaborated an study of the collaborative tagging system Delicious, where was able to discover regularities in user activity, tag frequencies, kings of tags used and bursts of popularity in bookmarking. Af- ter, the authors present a dynamic model of collabora- tive tagging that predicts these stable patterns and re- lates them to imitation and shared knowledge. Noll and Meinel  proposed a new approach to personalized Web searches based on collaboration and information sharing about Web documents. The proposal use social
In the field of data analytics grouping of like documents in textual data is a serious problem. A lot of work has been done in this field and many algorithms have purposed. One of them is a category of algorithms which firstly group the documents on the basis of similarity and then assign the meaningful labels to those groups. Description first clustering algorithm belong to the category in which the meaningful description is deduced first and then relevant documents are assigned to that description. LINGO (Label Induction Grouping Algorithm) is the algorithm of description first clustering category which is used for the automatic grouping of documents obtained from search results. It uses LSI (Latent Semantic Indexing); an IR (InformationRetrieval) technique for induction of meaningful labels for clusters and VSM (Vector Space Model) for cluster content discovery. In this paper we present the LINGO while it is using LSI during cluster label induction and cluster content discovery phase. Finally, we compare results obtained from the said algorithm while it uses VSM and Latent semantic analysis during cluster content discovery phase.
In this paper, we present PerTOHS theory for on-line handwriting segmentation inspired from the human perceptual system, based on the fact that handwriting is composed by a sequence of basic features approximated by perceptual codes. The elementary perceptual codes EPCs correspond to: Valley, Left oblique shaft, Shaft, Right oblique shaft. The GPCs are :Valley, Left oblique shaft, Shaft, Right oblique shaft, Occlusion, Left half opening occlusion, Right half opening occlusion, Up half opening occlusion, Down half opening occlusion, Ain. We have reviewed the filed of handwriting, theories of visual perception and reading/writing models, which are also on the base of our proposed theory. We detail the PerTOHS theory and its architecture based on the Beta-elliptic model for the generation of handwriting script and the perceptual encoding system for perceptual codes detection. We use the fuzzy set theory to detect EPCs and genetic algorithms to detect GPCs. We validate PerTOHS theory on on-line handwriting segmentaion and reccognition. We use MAYASTROUN database containing digits, letters, words in Arabic and Western language, Arabic texts, and IRONOFF database. The encouraging achieved results are similar to those produced by the human perceptual system during writing process and helpful for the analysis and the segmentation handwriting problems and allow an important reduction for the initial acquired information. A promising data reduction rate around 96% is obtained. Our perspectives are: using the generated GPCs as input for an on-line multilingual handwriting recognition system, and the development of an interactive user interface for handwriting learning using perceptual codes.
Abstract —With the advancement of computer technology, computer simulation in the field of education are more realistic and more effective. The definition of simulation is to create a virtual environment that accurately and real experiences to improve the individual. So Simulation Based Training is the ability to improve, replace, create or manage a real experience and training in a virtual mode. Simulation Based Training also provides large amounts of information to learn, so use data mining techniques to process information in the case of education can be very useful. So here we used data mining to examine the impact of simulation-based training. The database created in cooperation with relevant institutions, including 17 features. To study the effect of selected features, LDA method and Pearson's correlation coefficient was used along with genetic algorithm. Then we use fuzzy clustering to produce fuzzy system and improved it using Neural Networks. The results showed that the proposed method with reduced dimensions have 3% better than other methods.
In addition, this model avoids the variability in the analysis of newborn conditions provided by differ- ent health professionals, which could yield inequali- ties in the treatment. Besides, the fuzzymodel is very simple and implies in low computational expenses, making it possible an easy and inexpensive imple- mentation, features that have an important role in developing and poor countries. In cities where there are no experts available, the model can help under- standing and evaluating the risk of neonatal death based only on information regarding the gestational age and birth weight. This is available even in very modest conditions.
Diagnostic interplanetary network gateway (DING) protocol was developed to introduce monitoring capabilities in DTN technology . DING uses a subscription-basedmodel for information distribution to cope the connectivity of DTN. The goal of this work is to implement mechanisms for network monitoring on Bundle and lower layers that are not addressed by available mechanisms. The authors set up a DTN2 with 3 nodes, involving a several configurations steps. They evaluated this testbed switching DTN traffic between nodes using variable connectivity and bundle sizes. After this setup work, the next step is the implementation of network monitoring capabilities that would allow observer the status of the network and network traffic exchange at the DTN level. In original DTN2 implementation, no network monitoring functionality is built in. On this proposal, the nodes produce several log files created by shell scripts created over DTN2. Several parameters are considered because all of them are important for describing the status of the network. These parameters are the following: DTN nodes uptime, that shows whether a DTN node is running or not; Bundle Traffic, that shows a number of data related to Bundle traffic, such as how many Bundles may be received by a node and forward, how many are delivered, deleted or expired; and link statistics, that describes the status of each link between nodes, showing the speed of data transmission and the total of transferred data.
In the vector space model text is represented by a vector of terms. The definition of a term is not inherent in the model, but terms are typically words and phrases. If words are chosen as terms, then every word in the vocabulary becomes an independent dimension in a very high dimensional vector space. Any text can then be represented by a vector in this high dimensional space. If a term belongs to a text, it gets a non-zero value in the text-vector along the dimension corresponding to the term. Since any text contains a limited set of terms (the vocabulary can be millions of terms), most text vectors are very sparse. Most vector based systems operate in the positive quadrant of the vector space, i.e., no term is assigned a negative value. To assign a numeric score to a document for a query, the model measures the similarity between the query vector (since query is also just text and can be converted into a vector) and the document vector. The similarity between two vectors is once again not inherent in the model. Typically, the angle between two vectors is used as a measure of divergence between the vectors, and cosine of the angle is used as the numeric similarity cosine has the nice property that it is 1.0 for identical vectors and 0.0 for orthogonal vectors).
Due to the rapid growth of Information systems, the web becomes huger and people started using World Wide Web for everything. The increased use of World Wide Web makes the information management as a difficult task. In order to provide exact web information to the user, the information management system must have well organized and efficient techniques for information storage and retrieval. The web page contains many contents like images, video, files etc. The textual content of the web page is only considered to cluster and retrieval.To provide the exact information to the web user, the information system must have an efficient clustering technique. Clustering is a technique, which divides the web pages into groups called clusters so that web pages in each cluster are more analogous to each other than the pages from different clusters. Clustering techniques are used in several application areas such as pattern recognition, data mining, and machine learning, and so on.
Application of fuzzy techniques approach to model flexible system for the access to information on the WWW is also realized within the solved problem. The aim is to design the system that can represent and manages the vagueness and uncertainty, which is characteristic of the process of information searching and retrieval. When some specific information is searched, this point and click access paradigm is unpractical, and the effectiveness of the results strongly depends on the starting page. The definition of systems plays an important role that help users to automatically access information relevant to their needs , . The research is aimed at defining systems tolerant to imprecision and uncertainty in the elicitation of users' performances and able to learn them through an interactive and adaptive behaviour. The fuzzy technique approach is the definition basis of flexible system for locating and accessing information on the Web.
Joseph C. et al. (2006) described the development of a fuzzy neural network-based in-process mixed material- caused flash prediction (FNN-IPMFP) system for injection molding processes. The goal is to employ a fuzzy neural network to predict flash in injection molding operations when using recycled mixed plastics. Major processing parameters, such as injection speed, melt temperature, and holding pressure, were varied within a small range. The vibration signal data during the mold closing and injection filling stages was collected in real-time using an accelerometer sensor. The data was analyzed with neural networks and fuzzy reasoning algorithms, in conjunction with a multiple regression model, to obtain flash prediction threshold values under different parameter settings. The FNN-IPMFP system was shown to predict flash with 96.1% accuracy during the injection molding process.
The Web has emerged as a major player in information transfer and communication as a result of substantial increase in the volume of scholarly and business information published on the Web. This has underscored the need for more effective mechanisms for organization of information on the Web to facilitate efficient and effective retrieval than what is possible using the available search engines. Information and knowledge managers in organizations are increasingly facing a situation of both information overload on the one hand and increasing demand for filtered and relevant information on the other. Considering the sheer volume of information to be handled, it is generally conceded that solutions to handle such a situation should necessarily be technology-based and should make effective use of intelligent technologies. In fact the history of informationretrieval
Geographical InformationRetrieval (GIR) concerns the retrieval of information involving some kind of spatial awareness. Many documents contain some kind of spatial reference which may be important for IR. For example, to retrieve, rank and visualize search results based on a spatial dimension (e.g. “find me news stories about bush fires near Sidney”). Many challenges of geographic IR involve geographical references (geo-references) which systems need to recognized and treated properly. Documents contain geo-references expressed in multiple languages which may or may not be the same as the query language. For example, the city Cape Town (English) is also Kapstadt (German), Cidade do Cabo in Portuguese and Ciudad del Cabo (Spanish).