International Journal of Intelligent Systems and Applications @ijisa
Статьи журнала - International Journal of Intelligent Systems and Applications
Все статьи: 1187
Document summarization using textrank and semantic network
Статья научная
The research has implemented document summarizing system uses TextRank algorithms and Semantic Networks and Corpus Statistics. The use of TextRank allows extraction of the main phrases of a document that used as a sentence in the summary output. The TextRank consists of several processes, namely tokenization sentence, the establishment of a graph, the edge value calculation algorithms using Semantic Networks and Corpus Statistics, vertex value calculation, sorting vertex value, and the creation of a summary. Testing has done by calculating the recall, precision, and F-Score of the summary using methods ROUGE-N to measure the quality of the system output. The quality of the summaries influenced by the style of writing, the selection of words and symbols in the document, as well as the length of the summary output of the system. The largest value of the F-Score is 10% of the length ta of the document with the F-Score 0.1635 and 150 words with the F-Score 0.1623.
Бесплатно
Doppler ultrasound based non-invasive heart rate telemonitoring system for wellbeing assessment
Статья научная
Telemonitoring in the field of healthcare has vastly improved the quality of clinical diagnosis and disease prevention by providing timely medical consultation to people living in rural and remote areas. To monitor the health state of a patient certain vital physiological parameter like electrocardiogram (ECG), respiration rate, blood pressure, oxygen saturation, etc. are acquired and analyzed. Listening to the heart sounds (auscultation) is also a quick method to monitor the health state of the patient’s heart. In this paper, we propose the use of a portable Doppler ultrasound sensor for measuring the heart sounds reliably and to transmit the data for further clinical telemonitoring. We have developed an ultrasound-based hardware prototype which is non-invasive in nature and easy to operate. Its portability, high accuracy, low cost, and wireless nature make this device suitable for home-based self-diagnostic applications. The developed prototype was successfully able to capture both fundamental heart sounds S1 and S2 reliably and transfer the signal wirelessly to the LabVIEW-based monitoring and data logging unit. This unit extracts clinically useful health information like heart rate (HR), R-R interval and heart rate variability (HRV) using signal processing algorithms. Health information is then transmitted via the Internet to a distant hospital for further improved clinical diagnosis and consultancy. The prototype was validated on 40 healthy males in the age group of 25-35 years, and the results show an overall accuracy of 96.74% in HR detection when compared with an ECG sensor, a photoplethysmograph (PPG) sensor, a pulse oximeter device and manual auscultation.
Бесплатно
Dual Population Genetic Algorithm for Solving Constrained Optimization Problems
Статья научная
Dual Population Genetic Algorithm is an effective optimization algorithm that provides additional diversity to the main population. It addresses the premature convergence problem as well as the diversity problem associated with Genetic Algorithm. Thus it restricts their individuals to be trapped in the local optima. This paper proposes Dual Population Genetic Algorithm for solving Constrained Optimization Problems. A novel method based on maximum constrains satisfaction is applied as constrains handling technique and Dual Population Genetic Algorithm is used as meta-heuristic. This method is verified against 9 problems from Problem Definitions and Evaluation Criteria for the Congress on Evolutionary Computation 2006 Special Session on Constrained Real-Parameter Optimization problem set. The results are compared with existing algorithms such as Ant Bee Colony Algorithm, Differential Evolution Algorithm and Genetic Algorithm that have been used for solving same problem set. Analysis shows that this technique gives results close to optimum value but fails to obtain exact optimum solution. In future Dual Population Genetic Algorithm can produce more efficient solutions using alternative constrains handling technique.
Бесплатно
Dynamic Load Balancing using Graphics Processors
Статья научная
To get maximum performance on the many-core graphics processors, it is important to have an even balance of the workload so that all processing units contribute equally to the task at hand. This can be hard to achieve when the cost of a task is not known beforehand and when new sub-tasks are created dynamically during execution. Both the dynamic load balancing methods using Static task assignment and work stealing using deques are compared to see which one is more suited to the highly parallel world of graphics processors. They have been evaluated on the task of simulating a computer move against the human move, in the famous four in a row game. The experiments showed that synchronization can be very expensive, and those new methods which use graphics processor features wisely might be required.
Бесплатно
Dynamic Programming and Genetic Algorithm for Business Processes Optimisation
Статья научная
There are many business process modelling techniques, which allow to capture features of those processes, but graphical, diagrammatic models seems to be used most in companies and organizations. Although the modelling notations are more and more mature and can be used not only to visualise the process idea but also to implement it in the workflow solution and although modern software allows us to gather a lot of data for analysis purposes, there is still not much commercial used business process optimisation methods. In this paper the scheduling / optimisation method for automatic task scheduling in business processes models is described. The Petri Net model is used, but it can be easily applied to any other modelling notation, where the process is presented as a set of tasks, i.e. BPMN (Business Process Modelling Notation). The method uses Petri Nets’, business processes’ scalability and dynamic programming concept to reduce the necessary computations, by revising only those parts of the model, to which the change was applied.
Бесплатно
Статья научная
In this paper, a dynamic recurrent wavelet neural network observer and tracking control strategy is presented for a class of uncertain, nonaffine systems. In proposed scheme a dynamic recurrent wavelet network is used to design a nonlinear observer .Adaptation laws are developed for the online tuning of wavelet parameters. Based on the estimated states, a state feedback control law is derived to achieve the desired tracking performance. The stability of closed loop system and ultimate upper boundedness all closed loop signals is proven in Lyapunov sense. Effectiveness of proposed scheme is demonstrated through numerical simulation.
Бесплатно
Dynamic Vehicle Routing Problem: Solution by Ant Colony Optimization with Hybrid Immigrant Schemes
Статья научная
During past decades, several Meta-Heuristics were considered by researchers to solve Dynamic Vehicle Routing Problem.In this paper, Ant Colony Optimization integrated with Hybrid Immigrant Schemes methods are proposed for solving Dynamic Vehicle Routing Problem. Ant Colony Optimization with hybrid immigrant schemes methods namely HIACO-I, HIACO-II and HIACO-III focused on establishing the proper balance between intensification and diversification. The performance evaluation of the algorithms in which Random Immigrants and Elitism based Immigrants were hybridized in different proportions and added to Ant Colony Optimization algorithm showed that they had produced better results in many dynamic test cases generated from three Vehicle Routing Problem instances.
Бесплатно
Dynamic selection approach to overcome the demotivation of learners in a community learning system
Статья научная
Community of Practice (CoP) is a very rich concept for designing learning systems for adults in relation to their professional development. In particular, for community problem solving. Indeed, Communities of Practice are made up of people who engage in a process of collective learning in a shared domain. The members engage in joint activities and discussions, help each other, and share information. They build relationships that enable them to learn from each other. The most important condition for continuing to learn from a CoP is that the community should live and be active. However, one of the main factors of members demotivation to continue interacting through the CoP is the frequent receipt of a large number of aid requests related to problems that they might not be able to solve. Thing that may lead them to abandon the CoP. In an attempt to overcome this problem, we propose an approach for selecting a group of members who are the most appropriate to contribute to the resolution of a given problem. In this way, the aid request will be sent only to this group. Our approach consists of a static rules-based selection complemented with a dynamic selection based on the ability to solve previous similar problems through analysis of the history of interactions.
Бесплатно
E-mail Spam Filtering Using Adaptive Genetic Algorithm
Статья научная
Now a day’s everybody email inbox is full with spam mails. The problem with spam mails is that they are not malicious in nature so generally don’t get blocked with firewall or filters etc., however, they are unwanted mails received by any internet users. In 2012, more that 50% emails of the total emails were spam emails. In this paper, a genetic algorithm based method for spam email filtering is discussed with its advantages and dis-advantages. The results presented in the paper are promising and suggested that GA can be a good option in conjunction with other e-mail filtering techniques can provide more robust solution.
Бесплатно
E-reputation prediction model in online social networks
Статья научная
E-reputation management has become an important challenge for firms that try to improve their notoriety across the web and more specifically in social media. Indeed, the power of online communities to impact a brand’s image is undeniable and companies need a powerful system to measure their reputation as perceived by connected society. Moreover, they need to follow its variation and forecast its evolution to anticipate any impacting change. For this purpose we have implemented an Intelligent Reputation Measuring System (IRMS) that assesses reputation in online social networks on the basis of members’ activity and popularity. In this paper, we add a predictive module to IRMS that forecasts the evolution of reputation score using influence propagation algorithms.
Бесплатно
Economic Load Dispatch by Hybrid Swarm Intelligence Based Gravitational Search Algorithm
Статья научная
This paper presents a novel heuristic optimization method to solve complex economic load dispatch problem using a hybrid method based on particle swarm optimization (PSO) and gravitational search algorithm (GSA). This algorithm named as hybrid PSOGSA combines the social thinking feature in PSO with the local search capability of GSA. To analyze the performance of the PSOGSA algorithm it has been tested on four different standard test cases of different dimensions and complexity levels arising due to practical operating constraints. The obtained results are compared with recently reported methods. The comparison confirms the robustness and efficiency of the algorithm over other existing techniques.
Бесплатно
Efficiency Evaluation Metrics for Wireless Intelligent Sensors Applications
Статья научная
The metrology field has been progressed with the appearance of the wireless intelligent sensor systems providing more capabilities such as signal processing, remote multi-sensing fusion etc. This kind of devices is rapidly making their way into medical and industrial monitoring, collision avoidance, traffic control, automotive and others applications. However, numerous design challenges for wireless intelligent sensors systems are imposed to overcome the physical limitations in data traffic, such as system noise, real time communication, signal attenuation, response dynamics, power consumption, and effective conversion rates etc, especially for applications requiring specific performances. This paper analyzes the performance metrics of the mentioned sensing devices systems which stands for superior measurement, more accuracy and reliability. Study findings prescribe researchers, developers/ engineers and users to realizing an optimal sensing motes design strategy that offers operational advantages which can offer cost-effective solutions for an application.
Бесплатно
Efficient Data Clustering Algorithms: Improvements over Kmeans
Статья научная
This paper presents a new approach to overcome one of the most known disadvantages of the well-known Kmeans clustering algorithm. The problems of classical Kmeans are such as the problem of random initialization of prototypes and the requirement of predefined number of clusters in the dataset. Randomly initialized prototypes can often yield results to converge to local rather than global optimum. A better result of Kmeans may be obtained by running it many times to get satisfactory results. The proposed algorithms are based on a new novel definition of densities of data points which is based on the k-nearest neighbor method. By this definition we detect noise and outliers which affect Kmeans strongly, and obtained good initial prototypes from one run with automatic determination of K number of clusters. This algorithm is referred to as Efficient Initialization of Kmeans (EI-Kmeans). Still Kmeans algorithm used to cluster data with convex shapes, similar sizes, and densities. Thus we develop a new clustering algorithm called Efficient Data Clustering Algorithm (EDCA) that uses our new definition of densities of data points. The results show that the proposed algorithms improve the data clustering by Kmeans. EDCA is able to detect clusters with different non-convex shapes, different sizes and densities.
Бесплатно
Efficient and Fast Initialization Algorithm for K-means Clustering
Статья научная
The famous K-means clustering algorithm is sensitive to the selection of the initial centroids and may converge to a local minimum of the criterion function value. A new algorithm for initialization of the K-means clustering algorithm is presented. The proposed initial starting centroids procedure allows the K-means algorithm to converge to a “better” local minimum. Our algorithm shows that refined initial starting centroids indeed lead to improved solutions. A framework for implementing and testing various clustering algorithms is presented and used for developing and evaluating the algorithm.
Бесплатно
Статья научная
Classification is found to be an important field of research for many applications such as medical diagnosis, credit risk and fraud analysis, customer segregation, and business modeling. The main intention of classification is to predict the class labels for the unlabeled test samples using a labelled training set accurately. Several classification algorithms exist to classify the test samples based on the trained samples. However, they are not suitable for many real world applications since even a small performance degradation of classification algorithms may lead to substantial loss and crucial implications. In this paper, a simple classification method using the average weighted pattern score with attribute rank based feature selection has been proposed. Feature selection is carried out by computing the attribute score based ranking and the classification is performed using average weighted pattern computation. Experiments have been performed with 40 standard datasets and the results are compared with other classifiers. The outcome of the analysis shows the good performance of the proposed method with higher classification accuracy.
Бесплатно
Efficient clustering algorithm with enhanced cohesive quality clusters
Статья научная
Analyzing data is a challenging task nowadays because the size of data affects results of the analysis. This is because every application can generate data of massive amount. Clustering techniques are key techniques to analyze the massive amount of data. It is a simple way to group similar type data in clusters. The key examples of clustering algorithms are k-means, k-medoids, c-means, hierarchical and DBSCAN. The k-means and DBSCAN are the scalable algorithms but again it needs to be improved because massive data hampers the performance with respect to cluster quality and efficiency of these algorithms. For these algorithms, user intervention is needed to provide appropriate parameters as an input. For these reasons, this paper presents modified and efficient clustering algorithm. This enhances cluster’s quality and makes clusters more cohesive using domain knowledge, spectral analysis, and split-merge-refine techniques. Also, this algorithm takes care to minimizing empty clusters. So far no algorithm has integrated these all requirements that proposed algorithm does just as a single algorithm. It also automatically predicts the value of k and initial centroids to have minimum user intervention with the algorithm. The performance of this algorithm is compared with standard clustering algorithms on various small to large data sets. The comparison is with respect to a number of records and dimensions of data sets using clustering accuracy, running time, and various clusters validly measures. From the obtained results, it is proved that performance of proposed algorithm is increased with respect to efficiency and quality than the existing algorithms.
Бесплатно
Efficient intelligent framework for selection of initial cluster centers
Статья научная
At present majority of research is on cluster analysis which is based on information retrieval from data that portrays the objects and their association among them. When there is a talk on good cluster formation, then selection of an optimal cluster core or center is the necessary criteria. This is because an inefficient center may result in unpredicted outcomes. Hence, a sincere attempt had been made to offer few suggestions for discovering the near optimal cluster centers. We have looked at few versatile approaches of data clustering like K-Means, TLBOC, FEKM, FECA and MCKM which differs in their initial center selection procedure. They have been implemented on diverse data sets and their inter and intra cluster formation efficiency were tested using different validity indices. The clustering accuracy was also conducted using Rand index criteria. All the algorithms computational complexity was analyzed and finally their computation time was also recorded. As expected, mostly FECA and to some extend FEKM and MCKM confers better clustering results as compared to K-Means and TLBOC as the former ones manages to obtain near optimal cluster centers. More specifically, the accuracy percentage of FECA is higher than the other techniques however, it’s computational complexity and running time is moderately higher.
Бесплатно
Effort estimation of back-end part of software using chaotically modified genetic algorithm
Статья научная
The focus of Software Development Effort Estimation (SDEE) is to precisely predict the estimation of effort and time required for successfully developing a software project. From the past few years, data-intensive applications with a huge back-end part are contributing to the overall effort of projects. Therefore, it is becoming more important to add the back-end part to the SDEE process. This paper proposes an Evolutionary Learning (EL) based hybrid artificial neuron termed as dilation-erosion perceptron (DEP) framework from the mathematical morphology (MM) having its foundation in complete lattice theory (CLT) for solving the SDEE problem. In this work, we used the DEP (CMGA) model utilizing a chaotically modified genetic algorithm (CMGA) for the construction of DEP parameters. The proposed method uses the ER diagram artifacts such as aggregation, specialization, generalization, semantic integrity constraints, etc. for calculating the SDEE of back-end part of the business software. Furthermore, the proposed method was tested over two different datasets, one is existing and the other one is a self-developed dataset. The performance of the given method is then evaluated by three popular performance metrics, exhibiting better performance of the DEP (CMGA) model for solving the SDEE problems.
Бесплатно
Статья научная
This paper proposes a new, simple and user–friendly MATLAB built-in function, mathematical and Simulink models, to be used to early identify system level problems, to ensure that all design requirements are met, and, generally, to simplify Mechatronics motion control design process including; performance analysis and verification of a given electric DC machine, proper controller selection and verification for desired output speed or angle.
Бесплатно
Emotion Detection of Tweets in Indonesian Language using Non-Negative Matrix Factorization
Статья научная
Emotion detection is an application that is widely used in social media for industrial environment, health, and security problems. Twitter is ashort text messageknown as tweet. Based on content and purposes, the tweet can describes as information about a user’s emotion. Emotion detection by means oftweet, is a challenging problem because only a few features can be extracted. Getting features related to emotion is important at the first phase of extraction, so the appropriate features such as a hashtag, emoji, emoticon, and adjective terms are needed. We propose a new method for analyzing the linkages among features and reducedsemantically using Non-Negative Matrix Factorization (NMF). The dataset is taken from a Twitter application using Indonesian language with normalization of informal terms in advance. There are 764 tweets in corpus which have five emotions, i.e. happy (senang), angry (marah), fear (takut), sad (sedih), and surprise(terkejut). Then, the percentage of user’s emotion is computed by k-Nearest Neighbor(kNN) approach. Our proposed model achieves the problem of emotion detectionwhich is proved by the result near ground truth.
Бесплатно
