Slide background

جستجوي مقاله (جستجوی پیشرفته)

در این قسمت شما می توانید عنوان یا قسمتی از خلاصه مقاله مورد نظر خود را در کادر زیر وارد نمونه و لیست مقالات مرتبط را مشاهده نمایید

اخبار نشریه

آخرین شماره

No 14
Vol. 14 No. 4
Spring 2016
دانلود فایل |

آخرین مقالات منتشر شده

Data repositories contain sensitive information which must be protected from unauthorized access. Existing data mining techniques can be considered as a privacy threat to sensitive data. Association rule mining is one of the utmost data mining techniques which tries to cover relationships between seemingly unrelated data in a data base.. Association rule hiding is a research area in privacy preserving data mining (PPDM) which addresses a solution for hiding sensitive rules within the data problem. Many researches have be done in this area, but most of them focus on reducing undesired side effect of deleting sensitive association rules in static databases. However, in the age of big data, we confront with dynamic data bases with new data entrance at any time. So, most of existing techniques would not be practical and must be updated in order to be appropriate for these huge volume data bases. In this paper, data anonymization technique is used for association rule hiding, while parallelization and scalability features are also embedded in the proposed model, in order to speed up big data mining process. In this way, instead of removing some instances of an existing important association rule, generalization is used to anonymize items in appropriate level. So, if necessary, we can update important association rules based on the new data entrances. We have conducted some experiments using three datasets in order to evaluate performance of the proposed model in comparison with Max-Min2 and HSCRIL. Experimental results show that the information loss of the proposed model is less than existing researches in this area and this model can be executed in a parallel manner for less execution time
Golnar Assadat Afzali - Shahriar Mohammadi
DOI : 0
کلمات کلیدی : Big Data ، Association Rule ، Privacy Preserving ، Anonymization ، Data Mining
The problem of community detection has a long tradition in data mining area and has many challenging facet, especially when it comes to community detection in time-varying context. While recent studies argue the usability of social science disciplines for modern social network analysis, we present a novel dynamic community detection algorithm called COGNISON inspired mainly by social theories. To be specific, we take inspiration from prototype theory and cognitive consistency theory to recognize the best community for each member by formulating community detection algorithm by human analogy disciplines. COGNISON is placed in representative based algorithm category and hints to further fortify the pure mathematical approach to community detection with stabilized social science disciplines. The proposed model is able to determine the proper number of communities by high accuracy in both weighted and binary networks. Comparison with the state of art algorithms proposed for dynamic community discovery in real datasets shows higher performance of this method in different measures of Accuracy, NMI, and Entropy for detecting communities over times. Finally our approach motivates the application of human inspired models in dynamic community detection context and suggest the fruitfulness of the connection of community detection field and social science theories to each other.
Hamideh Sadat Cheraghchi - Ali Zakerolhossieni
DOI : 0
کلمات کلیدی : Social Network ، Clustering ، Cognitive Modeling ، Evolution
Although decreasing rate of death in developed countries because of Myocardial Infarction, it is turned to the leading cause of death in developing countries. Data mining approaches can be utilized to predict occurrence of Myocardial Infarction. Because of the side effects of using Angioplasty as main method for diagnosing Myocardial Infarction, presenting a method for diagnosing MI before occurrence seems really important. This study aim to investigate prediction models for Myocardial Infarction, by applying a feature selection model based on Wight by SVM and genetic algorithm. In our proposed method, for improving the performance of classification algorithm, a hybrid feature selection method is applied. At first stage of this method, the features are selected based on their weights, using weight by Support Vector Machine. At second stage, the selected features, are given to genetic algorithm for final selection. After selecting appropriate features, eight classification methods, include Sequential Minimal Optimization, REPTree, Multi-layer Perceptron, Random Forest, K-Nearest Neighbors and Bayesian Network, are applied to predict occurrence of Myocardial Infarction. Finally, the best accuracy of applied classification algorithms, have achieved by Multi-layer Perceptron and Sequential Minimal Optimization.
Hojjatollah Hamidi - Atefeh Daraei
DOI : 0
کلمات کلیدی : Artificial Neural Network ، Sequential Minimal Optimization ، REPTree ، Knowledge Discovery in Databases ، Myocardial Infarction
One of problems of OFDM systems, is the big value of peak to average power ratio. To reduce it, any attempt have been done amongst which, random phase updating is an important technique. In contrast to paper, since power variance is computable before IFFT block, the complexity of this method would be less than other phase injection methods which could be an important factor. Another interesting capability of random phase updating technique is the possibility of applying the variance of threshold power. The operation of phase injection is repeated till the power variance reaches threshold power variance. However, this may be a considered as a disadvantage for random phase updating technique. The reason is that reaching the mentioned threshold may lead to possible system delay. In this paper, in order to solve the mentioned problem, DCT transform is applied on subcarrier outputs before phase injection. This leads to reduce the number of required carriers for reaching the threshold value which results in reducing system delay accordingly.
Babak Haji Bagher Naeeni
DOI : 0
کلمات کلیدی : Orthogonal Frequency Division Multiplexing (OFDM) ، Peak-to-Average Power Ratio (PAPR) ، Random Phase Updating ، Orthogonal Frequency Division Multiplexing (OFDM) ، Peak-to-Average Power Ratio (PAPR) ، Random Phase Updating ، Discrete Cosine Transform ، ،
In this paper, a novel filter is provided that estimates the states of any nonlinear system, both in the presence and absence of uncertainty with high accuracy. It is well understood that a robust filter design is a compromise between the robustness and the estimation accuracy. In fact, a robust filter is designed to obtain an accurate and suitable performance in presence of modelling errors.So in the absence of any unknown or time-varying uncertainties, the robust filter does not provide the desired performance. The new method provided in this paper, which is named hybrid robust cubature Kalman filter (CKF), is constructed by combining a traditional CKF and a novel robust CKF. The novel robust CKF is designed by merging a traditional CKF with an uncertainty estimator so that it can provide the desired performance in the presence of uncertainty. Since the presence of uncertainty results in a large innovation value, the hybrid robust CKF adapts itself according to the value of the normalized innovation. The CKF and robust CKF filters are run in parallel and at any time, a suitable decision is taken to choose the estimated state of either the CKF or the robust CKF as the final state estimation. To validate the performance of the proposed filters, two examples are given that demonstrate their promising performance.
Behrooz Safarinejadian - Mohsen Taher
DOI : 0
کلمات کلیدی : Uncertainty ، State Estimation ، Cubature Kalman Filter (CKF) ، Robust CKF ، Hybrid Robust CKF

ابتداقبلی12بعدیانتها مشاهده 1 تا 5 ( از 10 رکورد)