New Publications are available for Knowledge engineering techniques
http://dl-live.theiet.org
New Publications are available now online for this publication.
Please follow the links to view the publication.Network active management for load balacing based in a intelligent multi agent system
http://dl-live.theiet.org/content/conferences/10.1049/cp.2012.0889
More than 1/3 of the EU's power will need to be generated from Renewable Energy Sources (RES) in 2020. This implies a profound transformation of Europe's energy system, grid and markets. The integration of distributed and micro generation in the electricity grid will provide more options for Distribution System Operators (DSOs) to balance their grid areas on the Medium Voltage (MV) and Low Voltage (LV) levels, thereby reducing stress at the Transmission System Operator (TSO) level. At the same time, this will place new requirements on TSOs and DSOs, particularly in terms of operational security. This paper describes the impact of intermittent distributed generation (DG) on distribution grid operation and how load balancing can be used to reduce that impact and to take advantage of it. (4 pages)Reinforcement learning based ALOHA for multi-hop wireless sensor networks with informed receiving
http://dl-live.theiet.org/content/conferences/10.1049/cp.2012.0582
In this paper, an ALOHA based Medium Access Control (MAC) protocol (RL-ALOHA with Informed Receiving) is proposed for multi-hop Wireless Sensor Networks (WSNs), which overcomes the traditional problems of low throughput, while exploiting their advantages of simplicity, low computational complexity and overheads. Reinforcement Learning (RL) is implemented as an intelligent slot assignment strategy in order to avoid collisions with minimal additional overheads. To improve the energy efficiency, Informed Receiving (IR) and ping packets are applied to avoid idle listening and overhearing. The simulation results show that this approach significantly increases the energy efficiency, achieves over twice throughput of Slotted ALOHA and reduces the end-to-end delay. (6 pages)Towards evidence-based ontology for supporting systematic literature review
http://dl-live.theiet.org/content/conferences/10.1049/ic.2012.0022
Systematic Literature Review (SLR) has become an important software engineering research method but costs tremendous efforts. This paper proposes an approach to leverage on empirically evolved ontology to support automating key SLR activities. [Method]: First, we propose an ontology, SLRONT, built on SLR experiences and best practices as a groundwork to capture common terminologies and their relationships during SLR processes; second, we present an extended version of SLRONT, the COSONT and instantiate it with the knowledge and concepts extracted from structured abstracts. Case studies illustrate the details of applying it for supporting SLR steps. Results show that through using COSONT, we acquire the same conclusion compared with sheer manual works, but the efforts involved is significantly reduced. The approach of using ontology could effectively and efficiently support the conducting of systematic literature review.Mass detection in digital mammograms using gabor filter bank
http://dl-live.theiet.org/content/conferences/10.1049/cp.2012.0465
Digital Mammograms are currently the most effective imaging modality for early detection of breast cancer but the number of false negatives and false positives is high. Mass is one type of breast lesion and the detection of masses is highly challenged problem. Almost all methods that have been proposed so far suffer from high number of false positives and false negatives. In this paper, a method for detecting true masses is presented, especially, for the reduction of false positives and false negatives. The key idea of the proposal is the use of Gabor filter banks for extracting the most representative and discriminative local spatial textural properties of masses that are present in mammograms at different orientations and scales. The system is evaluated on 512 (256 normal+256 true mass) regions of interests (ROIs) extracted from digital mammograms of DDSM database. We performed experiments with Gabor filter banks having different numbers of orientations and scales to find the best parameter setting. Using a powerful feature selection technique and support vector machines (SVM) with 10-fold cross validation, we report to achieve Az = 0.995±0.011, the area under ROC. Comparison with state-of-the-art techniques suggests that the proposed system outperforms similar methods, which are based on texture description, and the difference is statistically significant. (6 pages)Adaptive sampling in context-aware systems: a machine learning approach
http://dl-live.theiet.org/content/conferences/10.1049/cp.2012.0608
As computing systems become ever more pervasive, there is an increasing need for them to understand and adapt to the state of the environment around them: that is, their context. This understanding comes with considerable reliance on a range of sensors. However, portable devices are also very constrained in terms of power, and hence the amount of sensing must be minimised. In this paper, we present a machine learning architecture for context awareness which is designed to balance the sampling rates (and hence energy consumption) of individual sensors with the significance of the input from that sensor. This significance is based on predictions of the likely next context. The architecture is implemented using a selected range of user contexts from a collected data set. Simulation results show reliable context identification results. The proposed architecture is shown to significantly reduce the energy requirements of the sensors with minimal loss of accuracy in context identification. (5 pages)MRI mammogram image classification using ID3 algorithm
http://dl-live.theiet.org/content/conferences/10.1049/cp.2012.0464
Breast cancer is one of the most common forms of cancer in women. In order to reduce the death rate , early detection of cancerous regions in mammogram images is needed. The existing system is not so accurate and it is time consuming. The Proposed system is mainly used for automatic segmentation of the mammogram images and classify them as benign,malignant or normal based on the decision tree ID3 algorithm. A hybrid method of data mining technique is used to predict the texture features which play a vital role in classification. The sensitivity, the specificity, positive prediction value and negative prediction value of the proposed algorithm accounts to 93.45% , 99.95%,94% and 98.5% which rates very high when compared to the existing algorithms. The size and the stages of the tumor is detected using the ellipsoid volume formula which is calculated over the segmented region. (5 pages)Learning based objective evaluation of image segmentation algorithms
http://dl-live.theiet.org/content/conferences/10.1049/cp.2012.0444
Image segmentation plays an important role in a broad range of applications and many image segmentation methods have been proposed, therefore it is necessary to be able to evaluate the performance of image segmentation algorithms objectively. In this paper we present a new fuzzy metric to evaluate the accuracy of image segmentation algorithms, based on the features of each segments using neural networks. The neural network after training can distinguish the similarity or dissimilarity of each pairs of segments and finally the segmentation algorithms accuracy have been computed by novel presented metric quantitatively. Our proposed method does not require a manually-segmented reference image for comparison therefore can be used for real-time evaluation and is sensitive to both oversegmentation and under-segmentation. Experimental results were obtained for a selection of images from Berkeley segmentation data set and demonstrated that it's a proper measure for comparing image segmentation algorithms. (6 pages)Unsupervised learning of maritime traffic patterns for anomaly detection
http://dl-live.theiet.org/content/conferences/10.1049/cp.2012.0414
Maritime anomaly detection requires an efficient representation and consistent knowledge of vessel behaviour. Automatic Identification System (AIS) data provides ships state vector and identity information that is here used to automatically derive knowledge of maritime traffic in an unsupervised way. The proposed approach only utilises AIS data, historical or real-time, and is aimed at incrementally learning motion patterns without any specific a priori contextual description. This can be applied to a single AIS terrestrial receiver, to regional networks or to global scale tracking. The maritime traffic representation underpins low- likelihood behaviour detection and supports enhanced Maritime Situational Awareness by providing a characterisation of vessels traffic. (5 pages)A new multi-criteria fuzzy logic transformer inrush restraint algorithm
http://dl-live.theiet.org/content/conferences/10.1049/cp.2012.0044
This paper presents a new multi-criteria stabilization algorithm of transformer differential protection. Proposed scheme bases on new criteria signals and appropriate operation thresholds. New algorithm employs fuzzy reasoning technique for better discrimination of inrush conditions. The developed stabilization algorithm has been tested with ATP-EMTP generated signals, proving to be reliable and much more sensitive than standard stabilization algorithms with crisp settings. (6 pages)Emergency control scheme for transient stability using synchronized measurement
http://dl-live.theiet.org/content/conferences/10.1049/cp.2012.0139
A single-machine-equivalent (SIME) based emergency control (EC) scheme that incorporates synchronized measurements is described in the paper. This EC scheme resolves the transient stability problems experienced when serious disturbance occurs on critical transmission lines. The control actions are based on modifying the impedance of a thyristor controlled series compensator (TCSC). The SIME approach is first used to evaluate the transient stability of the system and then a decision is made about the control actions needed to stabilize the system. During emergency conditions, a fast response time is very important and this requires a security guideline to be used in the decision making process. The guideline is developed by analyzing offline multiple fault scenarios using a supervised learning approach. This ensures appropriate control actions can be performed without compromising the response time required on a real system. Finally, the emergency control (EC) scheme is tested and verified using an idealized four machine system. (5 pages)Audio classification based on sparse coefficients
http://dl-live.theiet.org/content/conferences/10.1049/ic.2011.0153
Audio signal classification is usually done using conventional signal features such as mel-frequency cepstrum coefficients (MFCC), line spectral frequencies (LSF), and short time energy (STM). Learned dictionaries have been shown to have promising capability for creating sparse representation of a signal and hence have a potential to be used for the extraction of signal features. In this paper, we consider to use sparse features for audio classification from music and speech data. We use the K-SVD algorithm to learn separate dictionaries for the speech and music signals to represent their respective subspaces and use them to extract sparse features for each class of signals using Orthogonal Matching Pursuit (OMP). Based on these sparse features, Support Vector Machines (SVM) are used for speech and music classification. The same signals were also classified using SVM based on the conventional MFCC coefficients and the classification results were compared to those of sparse coefficients. It was found that at lower signal to noise ratio (SNR), sparse coefficients give far better signal classification results as compared to the MFCC based classification. (5 pages)Randomised forests for people detection
http://dl-live.theiet.org/content/conferences/10.1049/ic.2011.0148
People detection is an important task with applications in fields such as surveillance and human computer interaction. A popular approach to this problem is to train a classifier on a data set using a particular set of features. A great deal of empirical evidence suggests that edge features are particularly discriminative for this task. In this paper we explore the use of randomised forests (sometimes referred to as randomised decision forests) for people detection. A randomised forest classifier is trained for people detection with edge orientation features. These features capture information concerning the distribution of edges with specific orientations. The classifier is trained and tested on the INRIA person data set, and some results are presented. (5 pages)A study on network intrusion detection and prevention system current status and challenging issues
http://dl-live.theiet.org/content/conferences/10.1049/ic.2011.0075
A network based Intrusion Prevention System sits in-line on the network, monitoring the incoming packets based on certain prescribed rules and if any bad traffic is detected, the same is dropped in real-time. A signature based detection system was developed to perform TCP port scans, Trace route scan, ping scan and packet sniffing to monitor network. This paper is going to enhance the signature based system to monitor network traffic, creation of per-flow packet traces and adaptive learning of intrusion. The existing Hawkeye solutions are used for the network intrusion detection and prevention system. In this paper we have proposed new model which will combine the three technique such as Adaptive weighted sampling algorithm, packet count flow classifier and Adaptive learning algorithms to the existing system.Target word sense disambiguation system for Kannada language
http://dl-live.theiet.org/content/conferences/10.1049/ic.2011.0097
The process of identifying the correct sense of a word in a specific context is called as Word Sense Disambiguation (WSD). It is essential for communication in a natural language. It is motivated by its use in many crucial applications such as Information retrieval, Information extraction, Machine Translation, Part-of-Speech tagging etc. The aim of our research is to develop a WSD system for target words in Kannada language. This paper presents our preliminary work towards building target word sense disambiguation system for Kannada language. To the best of our knowledge, this is the first attempt towards building WSD system for Kannada. Our work is a mile stone for Kannada language processing activities. In the present work, we exploited the compound words clue and syntactic features in a local context for target word sense disambiguation. It is noticed that, the use of syntax will improve the performance of the WSD system. The Kannada Shallow parser has been used for syntactic analysis. The ambiguous target word is disambiguated using supervised learning techniques. The experiments are conducted using Naive Bayes classifier. We created Kannada Corpora for evaluating the experiments and the results are encouraging.Multi-vehicle convoy analysis based on ANPR data
http://dl-live.theiet.org/content/conferences/10.1049/ic.2011.0135
This paper focuses on the development and novel application of data mining techniques for convoy analysis of vehicles based on the automatic number plate recognition (ANPR) system. The amount of ANPR data captured daily by traffic cameras in the road networks is very substantial. Data mining techniques are commonly used to extract relevant information and to reduce the amount of data processing and storage. In this paper, we apply data clustering techniques to extract relevant traffic patterns from the ANPR data to detect and identify unusual patterns and irregular behaviour of multi-vehicle convoy activities. (5 pages)A recurrent quantum neural network model enhances the EEG signal for an improved brain-computer interface
http://dl-live.theiet.org/content/conferences/10.1049/ic.2011.0028
The brain-computer interface (BCI) technology is a means of communication that allows individuals with severe movement disability to communicate with external assistive devices using the electroencephalogram (EEG) or other brain signals. The human mind and mental processes are inherently quantum in nature. It is therefore logical to investigate the possibility of designing new approaches to Brain-computer interface (BCI) with the amalgamation of quantum and classical approaches. This paper presents an intelligent information processing paradigm to enhance the raw electroencephalogram (EEG) data. A Recurrent Quantum Neural Network (RQNN) model using a non linear Schrodinger Wave Equation (SWE) is proposed here to filter the Motor Imagery (MI) based EEG signal of the BCI user. It is shown that if the potential field of the SWE is excited by the raw EEG data using a self-organized learning scheme, then the probability density function (pdf) associated with the EEG signal is transferred to the probability amplitude function which is the response of the SWE. In this scheme, the EEG data is encoded in terms of a particle like wave packet which helps to recover the EEG signal by denoising the raw data. Thus the filtered EEG signal is a wave packet which glides along and moves like a particle. This denoised EEG signal is then fed as an input to the feature extractor to obtain the Hjorth features. These features are then used to train a Linear Discriminant Analysis (LDA) classifier. It is shown that the accuracy of the classifier output over the training and the evaluation datasets using the filtered EEG is enhanced compared to that using the raw EEG signal for six of the nine subjects with a fixed set of parameters for all the subjects. (6 pages)Mining important predictors of heart attack
http://dl-live.theiet.org/content/conferences/10.1049/ic.2011.0067
Risk of heart attack is a global issue. This paper attempts to mathematically model the influence of eleven predictors on the heart attack risk. The contribution of each predictor and the related risk of heart attack are obtained from a group of medical doctors. Total 300 such cases are structured in a 300 * 12 matrix to conduct the study. Using statistical data mining, significant joint predictors have been extracted and clinically validated. The study also measures the variations in interpretations among doctors.Development of an ambient intelligent environment to facilitate the modelling of 'well-being'
http://dl-live.theiet.org/content/conferences/10.1049/ic.2011.0035
The monitoring of individuals conditions in a purely medical sense, through the acquisition and interpretation of physiological data by a medical practitioner on a regular, personal basis, will not be a viable proposition in the future due to the projected numbers of individuals requiring such activity. The programme of work being developed in the University of Glamorgan defines a networked, wireless ambient data acquisition environment that serves as real world Ambient Intelligent Environment (AIE) test bed. Its objective is the classification and integration of the data into a knowledge based intelligent system to provide the mechanism of ubiquitous computing. This paper presents a framework in which those ideas can be applied and tested in a distributed architecture that will facilitate invasive observation and even a level of supervision through pervasive and intelligent supervisor interventions. Initial work that has been undertaken has attempted to infer the emotional state of an individual within a monitored ambient intelligent environment and these early findings are presented as an indicator of the potential of the architecture being developed. The knowledge abstraction mechanism and classification techniques focus on the use of fuzzy logic methodology where a non-intrusive intelligent learning and adaptation agent that could be embedded in Ambient Intelligent Environments will be discussed. This proposed agent learns and models the user behaviour in order to control the environment on their behalf with respect to his emotional state. In order to realise one of the main requirements of ambient intelligent systems, the agent was developed to act in an adaptive way where it will manage and control the environment on behalf of the user with respect to his emotional state as an attempt to understand the word "Well- being". It will also allow the rules to be adapted and extended online, assisting a life-long learning technique as the environmental conditions changes and the user behaviour adjust with it. (6 pages)Context-aware support for assistive systems and services
http://dl-live.theiet.org/content/conferences/10.1049/ic.2011.0032
An assistive living system that gives our elderly population the comfort and confidence necessary to remain independently in their own homes is essential for enhanced longevity. Ambient Assistive Living technology that provides intuitive and context-sensitive support presents researchers with additional challenges. This paper details the creation and development of an "Ambient Assistant" application augmenting and extending the open source framework OpenAAL with enhanced reasoning, intelligent monitoring of the person and decision-making capabilities. The aim is a complete framework with fully interoperable components such as multi-parameter sensors. (4 pages)Study on food safety ontology reasoning application based on Jena
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0764
This study constructs a food safety ontology which adopts OWL (Web Ontology Language) to represent food safety knowledge in a structural way in order to help knowledge requesters clearly understand food safety knowledge; subsequently designs some ontology reasoning rules for deducing food safety knowledge in order to share and reuse relevant food safety knowledge effectively; finally implement an ontology-based food safety knowledge reasoning mechanism. Results of this study facilitate the food safety knowledge storage, management and sharing to provide knowledge requesters with accurate and comprehensive food safety knowledge for problem solving and decision support.An improved method using kinematic features for action recognition
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0766
Human action recognition is a challenge problem in computer vision. In this paper, we propose an improved approach using kinematic features for action recognition. In this approach, we find the area that relates to action by a simple method, and select eight discriminative features derived from optical flow field to describe the dynamics of the field. The covariance matrix of the feature vectors is used to fuse the features and to serve as the feature descriptor. Multi-class SVM classifiers are then employed for action classification. Experiments are carried out on public datasets. We obtain a recognition rate of 97.66% SEG-ACA and 98.2% SEQ-ACA on KTH dataset, and 98.89% SEQ-ACA and 93.83% SEG-ACA on WEIZMANN dataset with leave-one-out test.An algorithm proposed for semi-supervised learning in cancer detection
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0487
Semi-supervised learning, a relatively new area in machine learning, represents a blend of supervised and unsupervised learning, and has the potential of reducing the need of expensive labelled data whenever only a small set of labelled examples is available. In this paper an algorithm for Semi Supervised learning for detecting Cancer is proposed. We use the few labelled data to train the SVM classifier with Gist-SVM. We enlarge the number of training examples with SVM-Naive Bayes classifiers. We used WBC dataset from UCI Machine learning depository for our proposed methodology.Dynamic peer-to-peer distributed document clustering and cluster summarization
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0478
The main objective of this paper is to provide cluster summarization of huge text document. Mining process includes the sharing of large scale amount of data from various sources, which gets concluded at the mined data. In distributed data mining, adopting aflat node distribution model can affect scalability, modularity, flexibility which are being overcome by using dynamic peer to peer document clustering and cluster summarization. The Dynamic P2P document clustering and cluster summarization (DP2PCS) architecture is based upon bonus words and stigma words. For document clustering applications, the system summarizes the distributed document clusters using a distributed key-phrase extraction algorithm, thus providing interpretation of the clusters. Document summarization is used for fast information retrieval in less time. Compared to existing system the dynamic nature of proposed system facilitates a scalable cluster wherein the peers may join or leave the group at will. The summarization process on an average reduces the original documents content by 63 percentage based on the word count.Effect of mining educational data to improve adaptation of learning in e-learning system
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0499
The Apply of data mining technique in to domain specific applications is being drastically improved more and more by the researchers to improve the best decision making recommendation systems like e-learning ,e-commerce and etc.,In this research work we discuss how the datamining techniques will improve the learning style adaptation, learning content organisation and learning objects recommendations based on the instantaneous data collected through the web based learning management system like moodle. This paper focus the techniques like data classification and clustering techniques to predict the learning style of the peer learners based on the activities they have completed in the teaching learning activity of a particular course. The Experimental results shows the clear snopshot of the analysis.Adaptive congestion avoidance scheme based on reinforcement learning for wireless sensor network
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0664
Energy efficiency and QoS-aware are the key issues of wireless sensor network (WSN). In this paper, we proposed a congestion avoidance scheme devoting to efficient use of energy and adaptive maintain well QoS quality by self-adapt routing. Because it is difficult to obtain the state of network energy and QoS in a practical condition, we are motivated to utilize reinforcement learning to obtain the routing strategy in multi-path communication of WSN. We extend the R-learning algorithm to solve the difficulty of the nodes obtaining the network's status information. We compare the proposed scheme to other congestion avoidance protocols, such as CR. The simulation results show that the performance of our schemes is prior to existing ones.A comparative study of information extraction tools used for biological database
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0492
The Internet presents a huge amount of biological data. It is difficult to extract relevant data from various sources. Therefore, the availability of robust, flexible Information Extraction (IE) systems and tool that transform the Web pages into program-friendly structures such as a relational database. This paper made a study on information extraction tools. Which can be used for Biological databases. The tools have been classified based on four categories such as tools for Manually constructed Information Extraction, Supervised Wrapper Induction System, Semi supervised Information Extraction Systems, Unsupervised Information extraction Systems. Finally we made a comparative study on the Information Extraction tools used for Biological database based on the technique used such as scan Pass, Extraction Rules Type, Features used, Learning Algorithm and Tokenization Schemes.An improved ensemble learning method based on SVR
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.1453
As a universal learning method, Support Vector Regression (SVR) has strong generalization ability and can perfectly solve some practical problems, such as small samples, nonlinear, high dimension and so on. However, the prediction accuracy of a single SVR is limited. With the help of ensemble learning, the sample prediction accuracy of SVR can be effectively improved. In ensemble learning, the construction method of training samples is a key. The larger difference between the training sample sub-sets leads to the stronger generalization ability of SVR. In this work, an improved method of sample construction is proposed to increase the differences between training sample sets. The proposed method divides the samples into several sub-categories by clustering algorithm. Each sub-category adds the samples closing to the clustering center from other sub-categories to form a new training sample set. The experiment results demonstrate that the proposed improved method has less number of iterations and higher predict accuracy as compared to the method with random sampling.Study on food safety emergency topic detection model based on semantics
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.1438
A new method of self-adaptive emergency topic detection model based on semantics is proposed in this paper. We apply the CHI_LDA method to establish the model for the news topics and reports, so as to realize the topic modeling in the semantic feature space. It can resolve the problems of high dimension and sparseness in the feature space and semantic relevance, and improve the time efficiency for LDA method to realize the semantic mapping of the feature space. We improve the traditional Single-pass incremental clustering algorithm by optimizing the updating strategy of the topic model. Meanwhile we establish the topic detector combined with the news topic timing characteristics, and realize the self-adaptive learning of the topic model so as to track the dynamic changes in topic. Experimental results indicate that this method of topic detection has a better performance; it can further improve the effect of topic detection.Modeling of the Chinese driver's braking behavior in the simulated traffic scene of rear-end collision avoidance
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.1383
Drivers who exceed the speed limit and cut in aggressively are the majority in China. Their behaviors inevitably send other drivers into an execution cycle of irrational braking frequently in the complex traffic flow. So the main cause of severe traffic accidents and great casualties is Chinese drivers' raking behaviors. In this paper, Neural Network model of driver's braking behavior characteristics during ending-collision avoidance was set up. The driving behavior test under the scene of the simulated traffic environment was carried out on the fixed driving simulator (ADSL) improved by the State Key Laboratory of Automobile Dynamic Simulation of Jilin University. Neural Network GUI was set by Matlab to select the appropriate network structure and learning arithmetic, and finally the driver's braking behavior model was built up for the simulated ending-collision avoidance scene. It can be concluded that the Neural Network model can simulate driver's braking behavior during ending-collision avoidance exactly, and in our future researches this model with other simulated driving behaviors will be used to evaluate driving reliability on traffic safety.The research on the application of rough set analysis in the strategy of the road transport energy conservation and emission reduction
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.1412
It can effectively improve the accuracy of decision-making which data mining technique is applied to both the method and implementation of transportation energy saving. This paper presents a method of rough set analysis to reduce the project of fuel limit verification which is in the access system of Chinese road transport vehicles fuel consumption, and gets the actual verification of the application in Guangdong Province.Artificial intelligence technologies in business and engineering
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0486
Artificial intelligence (AI) is making its way back into the mainstream of corporate technology, this time at the core of business systems which are providing competitive advantage in all sorts of industries, including electronics, manufacturing, marketing, human resource, financial services software, medicine, entertainment, engineering and communications. Designed to leverage the capabilities of humans rather than replace them, today's AI technology enables an extraordinary array of applications that forge new connections among people, computers, knowledge, and the physical world. Some AI enabled applications are information distribution and retrieval, database mining, product design, manufacturing, inspection, training, user support, surgical planning, resource scheduling, and complex resource management. AI technologies help enterprises reduce latency in making business decisions, minimize fraud and enhance revenue opportunities.Study on cooperation of urban traffic control and route guidance based on fuzzy theory
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.1396
In this paper, first, by comparing the relative increment of the traffic flow with the occupancy, we can judge whether the road traffic flow is in the congestion formative stage. Then the road congestion is selected as the input of the fuzzy control, while the green light extension time is selected as the output of the fuzzy control. Next, in order to provide the reasonable input data of the fuzzy control, the detection coil is increased to two in every section under the premise of the data acquisition technology of the SCOOT system. Then the membership functions, the fuzzy control rules, the fuzzy reasoning and the defuzzification are identified by the MATLAB simulation software. Finally, according to the characteristic surface of the input/output between the road congestion and the green light extension time, this control system apparently meets the actual situation. Therefore, before the congestion, the cooperation of the urban traffic flow guidance system and the traffic flow control system is realized by the fuzzy theory.A novel adaptation approach for electromagnetic device optimization
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0073
In order to carry out a successful case adaptation in our case-based reasoning system for EM device design, we make use of semantic networks to organize related domain knowledge, and then construct a rule system as an inference engine which is based on the network. Based on the rule system, a novel adaptation algorithm is proposed to derive a new device case from an induction motor case-base with high dimensionality.A survey on information extraction using entity relation based methods
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0491
Information extraction (IE) is the process of extracting the essential elements from structured text or knowledge from unstructured text by identifying references to named entities as well as stated relationships between such entities. IE systems can be used to directly untangle abstract knowledge from a text corpus, or to extract concrete information from a set of documents which can further be analyzed with traditional data-mining techniques to discover more general patterns and to identify the relationship between the entities in the Structured Text. It is required to discuss the various Entity Relation Extraction Methods and accuracy of the Extraction process. This paper presents the comparison of different entity relation extraction methods between two entities.Frequency spectrum access mechanism of cognitive radio based on spectum prediction
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0730
In this thesis, a frequency spectrum access mechanism, in which future frequency spectrum state is predicted, is brought forward according to the fact that frequency spectrum detection error of Cognitive Radio is inevitable. That mechanism on support vector machine (SVM) method could evaluate the probability density of idle state duration and occupied state duration of authorized users in given channel. Therefore, it could assess and amend the result of frequency spectrum real-time detection , and could decrease the impact of false alarm probability and misjudgment probability on cognition system and authorized system. The result of simulation shows that the mechanism could obviously decrease the interference of detection error to authorized system, and could apparently reduce the lost of frequency spectrum access chances of cognition system caused by detection error.Traffic prediction model for cognitive networks
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.1431
As cognitive networks become so booming, many traditional network utilities must be reconsidered owing to uncertain and complicated changes after spectrum decision. It is a challenge for nodes to predict unknown network traffic precisely combined with spectrum characteristics. In this paper, we present a Relevance Vector Machine (RVM) based traffic prediction model. Based on the judgment of spectrum and wireless environments characteristics, networks traffic can be predicted with periodical samples training to form a close loop feedback. Simulation results for our model are presented and compared to Least Square Support Vector Machine (LS-SVM) scheme, and the simulation results show that the RVM solution improved prediction accuracy up to 60% at most.Semantic-based scene image classification
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.1446
As more and more people share and search the images on the Internet, there has strong demand to automatically classify the images and mimic human visual perception. This paper proposes a new semantic based classification method for scene images. It efficiently classifies the image with the semantic model of the scene. And it builds a 4-dimension object semantic histogram based on the low-level features of the major objects. The SVM is used for the multi-class classifier in order to get the scene semantics. The experiments show that this algorithm is very efficient and effective with a high accuracy level for scene images.Web questionnaire validation and vendor selection using adaptive neuro fuzzy inference system
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0476
Framing the questionnaire for supplier evaluation and selection is a serious problem in the automobile industry. Particularly, in an e-procurement scenario, the instrument has to be completely tested and verified before publishing the questionnaire for supplier evaluation. Therefore this paper presents an approach for validating the questionnaire using Factor analysis from the data set obtained from the manual records. The factors chosen for supplier evaluation are Cost, Quality, Service, Relationship, Organization and Past relationship. Initially, the 48 questions were confined to the questionnaire which was then confined into 9 components by Factor analysis. Finally the Adaptive Neuro Fuzzy Inference System was trained and tested to rate the supplier using the data collected from the e-procurement system.Development of an effective condition monitoring system for AC point machines
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0584
This paper develops a new approach for the fault detection and diagnosis of AC point machines. The paper shows that electrical active power can be used as a parameter for condition monitoring systems for AC point machines. The methodology proposed in this paper utilises Wavelet Transforms and Support Vector Machines. It was found that this method can detect and diagnose faults to a high degree of accuracy. (6 pages)A fall detection algorithm based on pattern recognition and human posture analysis
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0790
Detecting fall is a particular important task in security monitoring and healthcare applications of sensor networks. However traditional approaches suffer from either a high false positive rate or high false negative rate, especially when the collected sensor data are unbalanced. Therefore, there is a lack of tradeoff between false alarms and misses for many traditional data mining methods to be applied. To solve this problem a novel fall detection algorithm based on pattern recognition and human posture analysis is presented in this paper. It firstly extracts thirty temporal features from the original data traces for different length adaptation of samples, and then exploits Hidden Markov Model (HMM) to filter the noisy character data and reduce the dimension of feature vectors. After that, it performs a closer classification with one-class Support Vector Machine (OCSVM) to filter the high false positive samples, and finally applies posture analysis to counteract the effects of high false negative samples until a satisfying accuracy is achieved. Simulation with real data demonstrates that the proposed algorithm outperforms other existing approaches.System and methodology for unknown malware attack
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0475
Intrusion Detection Prevention Systems are increasingly a key part of systems defense. Various approaches to Intrusion Detection prevention are currently being used, but they are relatively ineffective. Artificial Intelligence plays a driving role in security services. This paper proposes a dynamic model Intelligent Intrusion Detection System, based on specific AI approach for intrusion detection. The techniques that are being investigated includes neural networks and fuzzy logic with network profiling, that uses simple data mining techniques to process the network data. The proposed system is a hybrid system that combines anomaly, misuse and host based detection.Internet tourism scene classification with multi-feature fusion and transfer learning
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0768
This paper proposes an internet tourism scene classification algorithm, named multi-feature fusion with transfer learning, which utilizes unlabeled auxiliary data to facilitate image classification. Firstly, we do the SURF extraction and MRHM analysis for the training data separately, in which the training data set as combined with labeled images and unlabeled auxiliary images. Then we compute the target feature vector for each image by merging the extended SURF descriptor and MRHM feature. Finally, we train the SVM classifier scene classification. Due to the capability of transferring knowledge, the proposed algorithm can effectively address insufficient training data problem for image classification. Experiments are conducted on a Beijing tourism scene dataset to evaluate the performance of our proposed algorithm. The experimental results are encouraging and promising.The application of an improved BP artificial neural network in distributed data mining
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0314
Distributed data mining (DDM) technique is becoming necessary for large and multi-scenario datasets requiring resource, which are heterogeneous and distributed. In this paper, we focused on distributed data of tax and the artificial neural network (ANN). Here we used an improved ANN to make a classification of tax data, first we use a set of sample data to train the improved ANN, so it could get a group of weights and then we could use the weights to make classification. From experiments we know it could reduce the training time and improve the classification accuracy. (5 pages)Multiple view learning based on tabular data
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0861
Comparing with single-view learning algorithm, multi-view learning algorithm has more powerful classification performance. However, multi-view learning algorithm needs multiple source patterns. Features of those multiple source information must satisfy with some independent conditions. In most real world case, it is easier for us to gain single source patterns. So it will be necessary for us to design multi-view learning algorithms that are on the basis of single source patterns. In our previous research, we proposed a multi-view learning algorithm which is named MultiV-MHKS and found that MultiV-MHKS can efficiently improve the recognition rate in multi-view learning. In the paper, on the basis of MultiV-MHKS, we propose a novel classification method named MultiV-TMHKS, which adopts the tabularized data technique to matrixize single source patterns. By this multiviewization approach we can gain different kinds of matrixes that are used in different views and then design proper sub-classifiers in corresponding views. We come up with a new matrixizing method for multiple view learning which is based on single source patterns.Analysis of the effect of the light rail on the airport passengers transfer
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.1386
For the large hub airports which are usually located far away from city center, the convenience of passenger transfer between city and airport is not only related to service quality of airport, but also to the important aspect of city traffic efficiency. This paper uses Beijing capital international airport as an example. It quantitatively analyzes the behavior of transfer of the airport passengers before and after the airport light rail bas built. First, twice airport passenger questionnaires have been used to collect the random samples of passenger transfer characteristic. Second, a support vector machine and Kullback-Leibler divergence have been used to analyze the overall changes of the passenger transfer samples and compare the distribution of the passengers transfer time before and after the airport light rail has been built, respectively. The results indicate that, the average departure time has been significantly shortened by the share of the light rail. But the potential of the light rail is still not fully brought into play, and there still has large improvement space for optimization of the Beijing capital international airport transfer.The ant colony optimization algorithm for web services composition on preference ontology
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.1455
The Optimization Algorithm for Web Services Composition on Preference Ontology (OAWSCP) is put forward. OAWSCP, which makes some improvements on primary ACO (Ant Colony Optimization), builds simulation model based on services composition, and sets multiple pheromones and pheromone weights to denote the preference to different properties of a service. The algorithm can also simulate the instability of the flow of services composition, and react according to the flow change of the services composition. The algorithm can also detect if the optimizing is converging to local optimization findings, and in this case the algorithm can take measures to change its direction, and as a result reduce the probability of the algorithm to converge to local optimization findings. In order to verify the feasibility of the algorithm, the paper also builds simulation application system. The result of the performance test proves that the algorithm is more effective than primary ACO.An intrusion detection scheme based on anomaly mining in Internet of Things
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.1014
Internet of things (IOT) is vulnerable to malicious attacks because of opening deployment and limited resources. It's heterogeneous and distributed characters make conventional intrusion detection methodologies hard to deploy. To overcome this problem, this paper shows an intrusion detection scheme based on the anomaly mining. The paper has two parts (i) in the first part an anomaly mining algorithm is developed to detect anomaly data of perception layer, (ii) in the second part a distributed intrusion detection scheme is designed based on the detected anomalies. Since not all anomalies are triggered by malicious intrusion, the intrusion semantic is analyzed to distinguish intrusion behaviors from anomalies. Finally our evaluation and analysis shows that our approach is accurate and extensible.Optimal placement of capacitor in radial distribution system using PSO
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0383
The problem of capacitor allocation in Electric Distribution Systems involves maximizing energy and peak power loss reductions by means of capacitor installations. This paper presents a novel approach using approximate reasoning to determine system candidate nodes in a distribution system for capacitor placement. The solution methodology has two parts: in part one the loss sensitivity factors are used to select the candidate locations for the capacitor placement and in part two the Particle Swarm Optimization Technique is used to identify the sizes of the capacitor for minimizing the energy loss cost and capacitor cost. The proposed method is applied to 15bus and 33bus radial distribution system.Towards a case-based computational model for the creative design of electromagnetic devices
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0082
In order to explore creativity in design, a computational model based on CBR (an approach to employing old experiences to solve new problems) and other soft computing techniques from machine learning, is proposed in this paper. The new model is able to address four challenging issues: generation of a design prototype from incomplete requirements; judgment and improvement of system performance given a sparse initial case base library; extraction of critical features from a given feature space; adaptation of retrieved previous solutions to similar problems for deriving a new solution to a given design task. The core principle employed by this model is that different knowledge from various level cases can be explicitly explored and integrated into a practical design process.Integrating rough clustering with fuzzy sets
http://dl-live.theiet.org/content/conferences/10.1049/cp.2011.0488
This paper presents the evolution and importance of clustering techniques, since clustering is unsupervised learning and there are many clustering methods in practice which results in which clustering scheme to be selected for our purpose .Here we take four clustering methodologies crisp Juzzy rough and rough fuzzy. These clustering methods have been implemented and its importance over one another is explained. And the suitable clustering method over these three has been identified for better perspective. The experiment results with the sample dataset illustrate the importance of clustering schemes.