Computational Collective Intelligence

ADELFE is a methodology dedicated to applications characterized by openness and the need of the system adaptation to an environment. It was proposed to guide the designer during the building of an Adaptive Multi-Agent System (AMAS). Given that designing such systems is not an easy task, this leads us to provide means to bring the AMAS design to a higher stage of automation and confidence thanks to Simulation. ADELFE 3.0 is a new version of ADELFE based on a Simulation Based Design approach in order to assist the designer of AMAS and make his task less difficult. This paper focuses on a practical example of building an AMAS using ADELFE 3.0.

The Flexible Job Shop Problem (FJSP) is an extension of classical job shop problem such that each operation can be processed on different machine and its processing time depends on the machine used. This paper proposes a new multi-agent model based on the meta-heuristic Chemical Reaction Optimization (CRO) to solve the FJSP in order to minimize the maximum completion time (makespan). Experiments are performed on benchmark instances proposed in the literature to evaluate the performance of our model.

The Job Shop problem with Time Lags (JSTL) is an important extension of the classical job shop scheduling problem, in that additional constraints of minimum and maximum time lags existing between two successive operations of the same job are added. The objective of this work is to present a distributed approach based on cooperative behaviour and a tabu search metaheuristic to finding the scheduling giving a minimum makespan. The proposed model is composed of two classes of agents: a Supervisor Agent, responsible for generating the initial solution and containing the Tabu Search core, and Resource_Scheduler Agents, which are responsible for moving several operations and satisfaction of some constraints. Good performances of our model are shown through experimental comparisons on benchmarks of the literature.

During last decades, the world has experienced a large number of economic crises, which were not confined to an individual economy, but affected directly, or in-directly almost every country all over the world. As a result, a number of international organizations, governments, and private sector institutions have begun to develop an early warning system as a monitoring system to detect the possibility of occurrence of an economic crisis in advance and to alert its users to take preventive actions. However, each of the systems addresses just one selected branch of economy, or particular spheres of company operation, and only a limited and small group of users apply them. Therefore this paper focuses on developing a conception of a model of a multiagent early warning system for crisis situations in economy. The system will include all branches of economy, and it may be used by any group of users. It will have the ability to predict unfavorable economic situations at a local, scattered, and integral level.

Increasing volume of traffic in urban areas causes great costs and has negative effect on citizens’ life and health. The main cause of decreasing traffic fluency is intersections. Many methods for increasing bandwidth of junctions exist, but they are still insufficient. At the same time intelligent, autonomous cars are being created, what opens up new possibilities for controlling traffic at intersections. In this article a new approach for crossing an isolated junction is proposed - cars are given total autonomy and to avoid collisions they have to change speed. Several methods for adjusting speed based on machine learning (ML) are described, including new methods combining different ML algorithms (hybrid methods). The approach and methods were tested using a specially designed platform MABICS. Conducted experiments revealed some deficiencies of the methods - ideas for addressing them are proposed. Results of experiments made it possible to verify the proposed idea as promising.

In recent years, the web has become an important software platform, with more and more applications becoming purely web-based. The agent technology needs to embrace these trends in order to remain relevant in the new era. In this paper, we present recent developments of our web-based multiagent middleware named Siebog. Siebog employs enterprise technologies on the server side in order to provide automatic agent load-balancing and fault-tolerance. On the client, it relies on HTML5 and related standards in order to run on a wide variety of hardware and software platforms. Now, with automatic clustering and state persistence, Siebog can support thousands of external devices hosting tens of thousands of client-side agents.

Coordination is one of the most important issues in order to reduce the damage caused by a crisis. To analyze the efficiency of a coordination plan, a BPMN (Business Process Modeling Notation version 2.0) model is usually used to capture the processes of activities and messages exchanged between the actors involved in a crisis, while an OCMAS (Organization Centered Multi-Agent System) model is used to represent the roles, their interactions and the organizational structures. In this paper, we describe a proposal that allows to perform an automatic transformation between BPMN and OCMAS models of the same coordination plan. The proposal is illustrated through a coordination plan of a tsunami evacuation.

Quality Management (QM) issues are together with production costs and delivery time one of the three main pillars of Lean Manufacturing. Although, Quality Operations Management should be supported by IT Manufacturing Execution Systems (MES), in practice it is very difficult to automate QM support on MES level because of its heterarchical and unpredictable nature. There is a lack of practical models that bind QM and MES. Authors try to fill this gap by proposed agent based MES architecture for QM support. This paper shows both concept of proposed architecture and its practical realisation on the example of automotive electronics device manufacturing.

This paper aim to outline an abstract template of theoretical framework for diagnosis and recovery of hierarchical multi-agent plans to deal with the plans execution exceptions. We propose a recovery policy based on the runtime diagnosis of resources and task failure. To this end, we take advantage of the ability to reason (using the formalism Hierarchical-Plan-Net-with-Synchronization, or

SHPlNet

) on the abstract tasks interaction and interference in order to localize the recovering region in the global plan.

Viral marketing, marketing techniques that use pre-existing social networks, has experienced a significant encouragement in the last years. In this scope, Twitter is the most studied social network in viral marketing and the rumor spread is a widely researched problem. This paper contributes with a survey of research works which study rumor diffusion in Twitter. Moreover, the most useful aspects of these works to build new multi-agent based simulations dealing with this interesting and complex problem are discussed. The main four research lines in rumor dissemination found and discussed in this paper are: exploratory data analysis, rumor detection, epidemiological modeling, and multi-agent based social simulation. The survey shows that the reproducibility in the specialized literature has to be considerably improved. Finally, a free and open-source simulation tool implementing several of the models considered in this survey is presented.

We present a methodology for identifying user communities on Twitter, by defining a number of similarity metrics based on their shared content, following relationships and interactions. We then introduce a novel method based on latent Dirichlet allocation to extract user clusters discussing interesting local topics and propose a methodology to eliminate trivial topics. In order to evaluate the methodology, we experiment with a real-world dataset created using the Twitter Searching API.

Abstract

Bipartite network is a performance of complex networks,The divided of unilateral node of bipartite network has important practical significance for the study of complex networks of community division. Based on the diffusion probability of information and modules ideas in the network,this paper presents a community divided clustering algorithm (IPS algorithm) for bipartite network unilateral nodes.The algorithm simulates the probability of information transfer in the network,through mutual support value between the nodes in network,selecting the max value as the basis for merger different communities.Follow the module of the definition for division after mapping the bipartite network nodes as a single department unilateral network.Finally,we use actual network test the performance of the algorithm.Experimental results show that,the algorithm can not only accurate divided the unilateral node of bipartite network,But also can get high quality community division.

In the current era, microblogging sites have completely changed the manner in which people communicate and share information. They give users the ability to communicate, interact, create conversations with each other and share information in real time about events, natural disasters, news, etc. On Twitter, users post messages called tweets. Tweets are short messages that do not exceed 140 characters. Due to this limitation, an individual tweet it’s rarely self-content. However, users cannot effectively understand or consume information.

In order, to make tweets understandable to a reader, it is therefore necessary to know their context. In fact, on Twitter, context can be derived from users interactions, content streams and friendship. Given that there are rich user interactions on Twitter. In this paper, we propose an approach for tweet contextualization task which combines different types of signals from social users interactions to provide automatically information that explains the tweet. To evaluate our approach, we construct a reference summary by asking assessors to manually select the most informative tweets as a summary. Our experimental results based on this editorial data set offers interesting results and ensure that context summaries contain adequate correlating information with the given tweet.

Many methods for measuring the semantic similarity between sentences have been proposed, particularly for English. These methods are considered restrictive as they usually do not take into account some semantic and syntactic-semantic knowledge like semantic predicate, thematic role and semantic class. Measuring the semantic similarity between sentences in Arabic is particularly a challenging task because of the complex linguistic structure of the Arabic language and given the lack of electronic resources such as syntactic-semantic knowledge and annotated corpora.

In this paper, we proposed a method for measuring Arabic sentences’ similarity based on automatic learning taking advantage of LMF standardized Arabic dictionaries, notably the syntactic-semantic knowledge that they contain. Furthermore, we evaluated our proposal with the cross validation method by using 690 pairs of sentences taken from old Arabic dictionaries designed for human use like Al-Wassit and Lissan-Al-Arab. The obtained results are very encouraging and show a good performance that approximates to human intuition.

Today’s leading businesses have understood the role of “social” into their everyday activity. Online social networks (OSN) and social media have melt and become an essential part of every firm’s concerns. Brand advocates are the new leading triggers for company’s success in online social networks and are responsible for the long term engagement between a firm and its customers. But what can it be said about this impressive crowd of customers that are gravitating around a certain brand advocacy or a certain community? Are they as responsive to a certain message as one might think? Are they really impressed by the advertising campaigns? Are they equally reacting to a certain comment or news? How they process the everyday grey knowledge that is circulating in OSN? In fact, how impressionable they are and which are the best ways a company can get to them?

Semi-supervised learning combines labeled and unlabeled examples in order to find better future predictions. Multi-view learning is another way to improve the prediction by combining training examples from more than one sources of data. In this paper, a semi-supervised multi-view learning approach is proposed for sentiment analysis in the Bulgarian language. Because there is little labeled data in Bulgarian, a second English view is also used. A genetic algorithm is applied for regression function learning. Based on the labeled examples and the agreement among the views on the unlabeled examples the error of the algorithm is optimized, striving after minimal regularized risk. The performance of the algorithm is compared to its supervised equivalent and shows an improvement of the prediction performance.

This paper presents the design phase of a work aiming at designing and developing a smart living device for seniors to assist them in their daily outdoor activities. We follow a participative design approach based on scenarios in order to design a socially-adapted device that will be useful to improve seniors’ life. To specify our system, we first provide an UML scenario metamodel to abstract all the concepts involved in our system collaborative functioning (interactions with stakeholders, environment …). This metamodel is used to generate different scenarios in order to better define future users ‘needs and the system requirements and notably its behavior (represented with BPMN and Petri Nets). A scenario generator has been implemented for that purpose. Finally, we show how to simulate and analyze those generated scenarios using process mining techniques.

The paper describes the problem of authority identification within web discussions solving using linear and nonlinear regression methods. The goal is to find an approximation of dependency of the authority value on variables representing parameters of the structure and particularly the content of selected web discussions. The approximation function can be used at first for computation of the authority value of a given discussant, at second, for discrimination of an authoritative discussant from non-authoritative contributors to the web discussion. This information is important for web users, who search for truthful and reliable information in the process of decision making about important things. The web users would like to be influenced by some credible professionals. The various regression methods were tested. The best solution was implemented in the Application for the Machine Authority Identification.

The semantic Web is an infrastructure that enables the interchange, the integration and the reasoning about information on the Web. In our annotation approach, we have proposed metadata of the semantic and fuzzy annotation in RDF to describe pages in a semantic Web environment. Our annotation of a page represents an enhancement of the first result of annotation done by the “Semantic Radar” Plug-in on the page. Now, we want to achieve a semantic and fuzzy interrogation of the annotated Web pages in order to improve the interrogation results for the domain experts. We propose in this paper a new editor named “

Sem-SPARQL Editor

”, which is based on the SPARQL language, to query the semantic and fuzzy annotations in RDF of pages.

This paper pioneers the avenue of enhancing a well-known paradigm in game playing, namely the use of

History

-based heuristics, with a totally-unrelated area of computer science, the field of Adaptive Data Structures (ADSs). It is a well-known fact that highly-regarded game playing strategies, such as alpha-beta search, benefit strongly from proper move ordering, and from this perspective, the

History

heuristic is, probably, one of the most acclaimed techniques used to achieve AI-based game playing. Recently, the authors of this present paper have shown that techniques derived from the field of ADSs, which are concerned with query optimization in a data structure, can be applied to move ordering in multi-player games. This was accomplished by ranking opponent threat levels. The work presented in this paper seeks to extend the utility of ADS-based techniques to two-player and multi-player games, through the development of a new move ordering strategy that incorporates the

historical

advantages of the moves. The resultant technique, the History-ADS heuristic, has been found to produce substantial (i.e, even up to 70%) savings in a variety of two-player and multi-player games, at varying ply depths, and at both initial and midgame board states. As far as we know, results of this nature have not been reported in the literature before.

Videos represent one of the most frequently used forms of multimedia applications. In addition to watching videos, people control slider bars of video players to find specific scenes and want detailed information on certain objects in scenes. However, it is difficult to support user interactions in current video formats because of a lack of metadata for facilitating such interactions. This paper proposes a text-based semantic video annotation system for interactive cooking videos to facilitate user interactions. The proposed annotation process includes three parts: the synchronization of recipes and corresponding cooking videos based on a caption-recipe alignment algorithm; the information extraction of food recipes based on lexico-syntactic patterns; and the semantic interconnection between recognized entities and web resources. The experimental results show that the proposed system is superior to existing alignment algorithms and effective in semantic cooking video annotation.

In this article is presented algorithm for obstacle detection and objects tracking in a railway crossing area. The object tracking is based on template matching and sum of absolute differences. The object tracking was implemented for better reliability of presented system. For optical flow estimation is used a modified Lucas-Kanade method. The results of proposed algorithm were verified in a real traffic scenarios consisted of two railway crossings in Czech Republic during 2013-14 under different environmental conditions.

This paper introduces a new approach for detecting and classifying baggage carried by human in images. The human region is modeled into several components such as head, body, foot and bag. This model uses the location information of baggage relative to human body. Features of each component is extracted. The features are then used to train boosting support vector machine (SVM) and mixture model over component. In experiment, our method achieves promising results in order to build automatic video surveillance system.

Association rule mining (ARM) is well-known issue in data mining. It is a combinatorial optimization problem purpose to extract the correlations between items in sizable data-sets. According to the literature study, bio-inspired prove their efficiency in term of time, memory and quality of generated rules. This paper investigates multi-population cooperative version of bat algorithm for association rule mining (BAT-ARM) named MPB-ARM which is based on bat inspired algorithm. The advantage of bat algorithm is the power combination between population-based algorithm and the local search, however, it more powerful in local search. The main factor to judge optimization algorithms is ensuring the interaction between global diverse exploration and local intensive exploitation. To maintain the diversity of bats, in our proposed approach, we introduce a cooperative master-slave strategy between the subpopulations. The experimental results shows that our proposal outperforms other bio-inspired algorithms already exist and cited in the literature including our previous work BAT-ARM.

Deep convolutional neural networks (DCNs) are increasingly being used with considerable success in image classification tasks trained over large datasets. However, such large datasets are not always available or affordable in many applications areas where we would like to apply DCNs, having only datasets of the order of a few thousands labelled images, acquired and annotated through lenghty and costly processes (such as in plant recognition, medical imaging, etc.). In such cases DCNs do not generally show competitive performance and one must resort to fine-tune networks that were costly pretrained with large generic datasets where there is no a-priori guarantee that they would work well in specialized domains. In this work we propose to train DCNs with a greedy layer-wise method, analogous to that used in unsupervised deep networks. We show how, for small datasets, this method outperforms DCNs which do not use pretrained models and results reported in the literature with other methods. Additionally, our method learns more interpretable and cleaner visual features. Our results are also competitive as compared with convolutional methods based on pretrained models when applied to general purpose datasets, and we obtain them with much smaller datasets (1.2 million vs. 10K images) at a fraction of the computational cost. We therefore consider this work a first milestone in our quest to successfully use DCNs for small specialized datasets.

In this paper we model a sequence of decisions via a simple majority voting game with some players possessing an unconditional or conditional veto. The players vote (yes, no or abstain) on each motion in an infinite sequence, where two rounds of voting take place on each motion. The form of an equilibrium with retaliation is introduced, together with necessary and sufficient conditions for an equilibrium in such a game. A theorem about the form of the equilibrium is proved: given that one veto player abstained in the first round, in the second round of voting on a motion: (1) a veto player should vote for the motion if the value of the

j

-th motion to the

i

-th player (measured relative to the status quo) is greater than

t

2

, veto if it is less than

t

1

and otherwise abstain; (2) a non-veto player should vote for the motion if the value of the

j

-th motion to the

i

-th player (measured relative to the status quo) is greater than

t

3

and otherwise abstain; (3) the thresholds

t

1

,

t

2

and

t

3

satisfy given conditions.

Sparse representations of signals based on learned dictionaries have drawn considerable interest in recent years. However, the design of dictionaries adapting well to a set of training signals is still a challenging problem. For this task, we propose a novel algorithm based on DC (Difference of Convex functions) programming and DCA (DC Algorithm). The efficiency of proposed algorithm will be demonstrated in image denoising application.

The concept of a large margin is central to support vector machines and it has recently been adapted and applied for nearest neighbour classification. In this paper, we suggest a modification of this method in order to be used for regression problems. The learning of a distance metric is performed by means of an evolutionary algorithm. Our technique allows the use of a set of prototypes with different distance metrics, which can increase the flexibility of the method especially for problems with a large number of instances. The proposed method is tested on a real world problem – the prediction of the corrosion resistance of some alloys containing titanium and molybdenum – and provides very good results.

Artificial immune system (AIS) algorithms have been successfully applied in the domain of supervised learning. The main objective of supervised learning algorithms is to generate a robust and generalized model that can work well not only on seen data (training data) but also predict well on unseen data (test data). One of the main issues with supervised learning approaches is model overfitting. Model overfitting occurs when there is insufficient training data, or training data is too simple to cover the structural complexity of the domain being modelled. In overfitting, the final model works well on training data because the model is specialized on training data but provides significantly inaccurate predictions on test data due to the model’s lack of generalization capabilities. In this paper, we propose a novel approach to address this model overfitting that is inspired by the processes of natural immune systems. Here, we propose that the issue of overfitting can be addressed by generating more data samples by analyzing existing scarce data. The proposed approach is tested on benchmarked datasets using two different classifiers, namely, artificial neural networks and C4.5 (decision tree algorithm).

Indoor localisation systems based on a Wi–Fi local area wireless technology bring constantly improving results. However, the whole localisation system may fail when one or more Access Point (AP) malfunctions. In this paper we present how to limit the number of observed APs and how to create a malfunction immune localisation method. The presented solutions are an ensemble of random forests with an additional malfunction detection system. The proposed solution reduces a growth of the localisation error to 4 percent for the floor detection inside a six floor building and 2 metres for the horizontal detection in case of a gross malfunction of an AP infrastructure. The system without proposed improvements may give the errors greater than 30 percent and 7 metres respectively in case of not detected changes in the AP’s infrastructure.

Unmanned Aerial Vehicles have been a growing field of study over the last few years. The use of unmanned systems require a strong human supervision of one or many human operators, responsible for monitoring the mission status and avoiding possible incidents that might alter the execution and success of the operation. The accelerated evolution of these systems is generating a high demand of qualified operators, which requires to redesign the training process to deal with it. This work aims to present an evaluation methodology for inexperienced users. A multi-UAV simulation environment is used to carry out an experiment focused on the extraction of performance profiles, which can be used to evaluate the behavior and learning process of the users. A set of performance metrics is designed to define the profile of a user, and those profiles are discriminated using clustering algorithms. The results are analyzed to extract behavioral patterns that distinguish the users in the experiment, allowing the identification and selection of potential expert operators.

Nowadays, institutions of higher education need to be informed about the opinion of their students about the services offered. They depend on this information to make strategic decisions. To accomplish this, many choose to hire consulting services companies that have traditional survey or focus group to obtain the required information; however, the bias that there is on this type of study, means that in some cases strategic decisions are not entirely reliable. Discussions of people in the Web 2.0 are a relevant source of information for this kind of organization due to the large amount of data that is posted every day. A case in point are the discussion forums, which made it possible for people to share their views freely on a specific topic. This is achieved thanks to comments made in publications; from which useful information can be extracted for use in other purposes. This paper will provide a prototype to exploit the information contained within the comments made in discussion forums. These sources of knowledge are often not processed for any purpose, however the proposed prototype allow us to extract relevant data from this reliable source to use as a knowledge base for a institution of higher education.

A growth of data published on the Web is still observed. Keyword-based search, used by most search engines, is a common way of information retrieval on the Web. Subsequently, keyword-based search may provide a huge amount of retrieved valueless information. This problem can be solved by Question Answering System (QAS, QA system). One of the challenging tasks for available QA systems is to understand the natural language questions correctly and deduce the precise meaning to retrieve accurate responses. A significant role of QA and an increasing number of them may cause a problem with selection the most suitable QA system. The general aim of this paper is to provide knowledge-based approach to QA system selection. It should ensure knowledge systematization and help users to find a proper solution that meets their needs.

As the amount of digital information is exploding in social, industry and scientific areas, MapReduce is a distributed computation framework, which has become widely adopted for analytics on large-scale data. Also, the idea which is used to solve the large-scale data problem by the use of approximation algorithms has become a very important solution in recent years. Especially for solving high-dimensional text data processing, semantic Web and search engine are required to pay attention to proximity searches and text relevance analysis. The difficulties of large-scale text processing mainly include its quick comparison and relevance judgment. In this paper, we propose an approximate bit string for approximation search method on MapReduce platform. Experiments exhibits excellent performance on efficiency effectiveness and scalability of the proposed algorithms.

This paper presents a new approach to performing linguistic summaries of graph datasets with the use of ontologies. Linguistic summarization is a well known data mining technique, aimed to discover patterns in data and present them in natural language. So far, this method has been applied only to relational databases. However amount of available graph datasets with associated ontologies is growing fast, hence we have investigated the problem of applying linguistic summaries in this scenario. As our first contribution, we propose to use an ontological class as subject of a summary, showing that its class taxonomy has to be used to properly select objects for summarization. Our second contribution is an extension to a summarizer, by analysis of set of ontological superclasses. We then propose extensions to quality measures

$$T_1$$

and

$$T_2$$

, measuring informativeness of a summary in the context of ontological class taxonomy. We also show that our approach can create more general summarizations (higher in class taxonomy). We verify our proposals by performing linguistic summarization on Semantic Web, which is a vast distributed graph dataset with several associated ontologies. We conclude the paper with showing the possibilities of future work.

Movie summarization focuses to obtain as much as possible of information as a shorter movie clip does, that but it keeps the content of the original and presents to the audience the faster way for understanding the movie. In this paper, we propose a co-occurrence characters network analysis for movie summarization based on discovery and analysis movie storytelling. Experiments on 17 movies in the Star War series, the Lord of the Ring series and Harry Porter series with more than 2000 minutes of movies play time and the evaluated results are compared to IMDb and IMSDb database. Our results show that proposed method has outperformed the conventional approaches in terms of the movie summarization rate.

The Proactive Context Aware Recommender Systems aim at combining a set of technologies and knowledge about the user context not only in order to deliver the most appropriate information to the user need at the right time but also to recommend it without a user query. In this paper, we propose a contextualized proactive multi-domain recommendation approach for mobile devices. Its objective is to efficiently recommend relevant items that match users’ personal interests at the right time without waiting for users to initiate any interaction. Our contribution is divided into two main areas: The modeling of a situational user profile and the definition of an aggregation frame for contextual dimensions combination.

User modeling is crucial aspect of personalized document retrieval systems. In this paper we propose to use ontology-based user profile while ontological structures are appropriate to represent dependencies between concepts in user profile. A method for clustering set of users is proposed. As a similarity measure between ontological profiles we present a novel approach using ontology alignment methods. To avoid “cold-start problem” we developed method for profile recommendation for a new user.

This contribution presents a new approach to the representation of user interests and preferences at information retrieval process on the Web. The adaptive user profile includes both interests given explicitly by the user, as a query, and also preferences expressed during relevance valuation process, so to express field independent translation between terminology used by the user and terminology accepted in some field of knowledge. Building, modifying, expanding (by semantically related terms) and using procedures for the profile are presented. Experiments concerning the profile, as a personalization mechanism of Web retrieval system, are presented and discussed.

We present probably the first formal theory of A. Smith’s Invisible Hand paradigm (ASIH). It proves that it is not only an idea, that is often conflicting with established governing methods, but something real, for which a formal theory can be built. This should allow for the creation of new tools for market analysis and prediction. It claims, that a market has another dimension of computational nature, which itself is a complete programmable computer; however quite different from a digital computer. There, unconscious meta-inference process of ASIH is spread on the platform of brains of agents and the structure of a market. This process is: unconscious, distributed, parallel, non-deterministic with properties of chaotic systems. A given thread of ASIH emerges spontaneously in certain market circumstances and can vanish when the market situation changes. Since this computer is made up of brains of agents, conclusions of this inference process affect the behavior of agents and therefore the behavior of the entire market. For a description of ASIH, a molecular model of computations must be used. Our theory shows that ASIH is much more “universal” than expected, and is not only restricted to market optimization and stabilization, but can also act as a “discoverer” of new technologies, (technical market optimization). Also rules of social behavior can be discovered (social market optimization). ASIH can be considered as Collective Intelligence (CI) of a market, similarly to Collective Intelligence of an ant hill.

This paper describes a comparative study of edge operators for the task of detecting dominant vanishing points in the image. Segmentation of line is required in order to detect the vanishing points. Three edge operators such as Sobel, Canny, and LoG (Laplacian of Gaussian) are used in order to compare that edge has influence on detecting the vanishing points. Most of line segments are obtained based on edge detection. The vanishing points are estimated by MSAC (m-estimator sample consensus) based algorithm. First, lines are extracted from edge images produced by edge operators. Second, vanishing points are obtained. The results of line segments and vanishing points detection are compared and discussed. The comparison is carried out based on the result of implementation on images with different buildings.

Allen temporal relations is a well-known formalism used for modeling and handling temporal data. This paper discusses an idea to introduce some flexibility in defining such relations between two fuzzy time intervals. The key concept of this approach is a fuzzy tolerance relation conveniently modeled. Tolerant Allen temporal relations are then defined using the dilated and the eroded intervals of the initial fuzzy time intervals. This extension of Allen relation is investigated for the purpose of temporal databases querying thanks to the language

TSQLf

introduced in our previous works.

To make use of collective intelligence of many autonomous self-interested agents, it is important to form a team that all the agents agree. Two-sided matching is one of the basic approaches to form a team that consists of agents from two disjoint agent groups. Traditional two-sided matching assumes that an agent has totally ordered preference list of agents to be paired with. However, it is unrealistic to have a totally ordered list for a large-scale two-sided matching problem. Therefore, two-sided matching with indifferences is proposed. It allows indifferences in the preference list of agents. Two-sided matching with indifferences has two important characters weakly stable and Pareto efficiency. In this paper, we propose a new integer programming based algorithm “nucleolus” for two-sided matching with indifferences. This algorithm propose the matching which satisfies weakly stable and Pareto efficiency.

The smart antennas are the antenna arrays with smart signal processing algorithms used to track and locate the antenna beam on the target. The idea of smart antennas is to use base station antenna patterns that are not fixed, but adapt to the current radio conditions. The DPA smart antenna may simplify the problem of design and implementation in comparison of adaptive or switched lobe method because of their capabilities of interference suppression, easy design and implementation. A simple DPA based smart antenna using dipole leg is proposed in this paper which makes it suitable for practical implementation without any compromise in its performance, thus avoiding the need for a high cost adaptive array. Side lobe magnitude and number of sidelobe are reduced by adding null point to the design of array signal in conventional Fourier series window method techniques. A control algorithm based on RBFNN technique is proposed to control the variations in shape and interference which provides the performance of DPA smart antenna equivalent to adaptive array technique. It effectively controls the directivity pattern variations without sacrificing the spectral performance of antenna array.

In this work, we compare three classifiers in terms of accuracy. The first is a copula-based Bayesian classifier based on elliptical and Archimedean copulas. The remaining two are Naive Bayes and Neural Networks. Such a comparison, particularly for the recently proposed Archimedean copula-based Bayesian classifiers, hasn’t been reported in the literature. The results show that copula-based Bayesian classifiers are a viable alternative to Neural Networks in terms of accuracy while keeping the models relatively simple.

Event Named entity Recognition (NER) is different from most past research on NER in Arabic texts. Most of the effort in named entity recognition focused on a specific domains and general classes especially the categories; Organization, Location and Person. In this work, we build a system for Event named entities annotation and recognition. To reach our goal we combined between linguistic resources and tools. Our method is fully automatic and aims to ameliorate the performance of our machine translation system.

An unstructured knowledge processing in integrated management information systems is increasingly becoming a major challenge, mainly due to the possibility to obtain better flexibility and competitiveness of the organization. However, the most prevailing phenomenon is a conflict in unstructured knowledge. As an example may serve opinions of users about a given product offered by online shops. Some users may have positive opinions; others negative ones, while some of them may not have any opinion about a given product. Therefore this paper focus on developing a consensus deriving method for resolving conflict of unstructured knowledge of text documents represented by Term Frequency Matrix in integrated management information system.