Frontiers in Health Informatics
Volume:2 Issue: 1, 2013

Breast cancer is one of the cancer that affect women the most. In order to reduce the morbidity and mortality rate due to breast cancer, the best way is to screen and diagnose it as early as possible. There are several ways to diagnose breast cancer; one of them is using mammography. In mammography, cancer or lesions are detected by validating the lesion density, size, shape and nature. Since density is one of the parameter of detecting lesions, it may be difficult for the radiologists to detect the lesions in dense breasts. This study is organized to determine the distribution and association of breast density with age and ethnicity. A cross sectional study was conducted in Department of Imaging, Country Height Health Sanctuary. All medical record of patients underwent medical check-up at the center from January 2007 to December 2009 were selected for this study. Standardized checklist was used to retrieve the data from available medical record of patients in the center. With the total number of 610 subjects, there were significant associations between breast density and age group. However, there were no significant association between breast density and ethnic groups. As conclusion Age has association with breast density while there are no significant association between breast density and ethnic groups1.

Feature Selection is simply defined as the process of identifying a small subset of highly predictive features out of a large set of candidates. However، in the presence of special conditions like facing imbalanced dataset، this is a rather different process. In this study we consider feature selection as a Reinforcement Learning problem and use the well-known method، i. e. Monte Carlo، to traverse the state space and select the best subset of features. Since، this is a pragmatic procedure to learn from the environment، it is highly adaptive to all of the conditions. Specifically، at first we consider the state space as a Markov Decision Process (MDP) and then introduce an optimal graph search to overcome the complexity of the problem of concern. This method initially explores the space of feature sets، and then exploits the accomplished experiments. According to the conditions of imbalanced datasets، the criterions that consider the circumstances of skewed classes are used for evaluation. The effectiveness and efficiency of our work is demonstrated through conducting several experiments compared with the other state-of-the-art methods1.

This paper focuses on the Delay/Fault-Tolerant Mobile Sensor Network for pervasive information gathering by a priority queue timer. We develop simple and efficient data delivery schemes tailored for DFT-MSN, which has several unique characteristics, such as sensor mobility, loose connectivity, fault tolerability, delay tolerability, and buffer limit. We first study two basic approaches, namely,direct transmission and flooding. We analyze their performance by using queuing theory and statistics. Based on the analytic results that show the trade-off between data delivery delay/ratio and transmission overhead, we introduce an optimized flooding scheme that minimizes transmission overhead in flooding. Then, we propose two simple and effective DFT-MSN data delivery schemes, Based on fault-tolerance and time. The waiting time for fault-tolerance scheme applies to messages that indicate the importance of the message. Decisions on the basis of a fault-tolerant transmission and eliminates waiting time The data is taken to minimize overflow. The decisions on message transmission and dropping are made based on fault tolerance for minimizing transmission overhead. The system parameters are carefully tuned on the basis of thorough analyses to optimize network performance. Extensive simulations are carried out for performance evaluation. Our results show that both schemes achieve ahigh message delivery ratio with acceptable delay1.

The value of learning rate and its change mechanisms is one of the issues in designing learning systems such as learning automata. In most cases a time-based reduction function is used to adjust the learning rate aim at reaching stability in training system. So the learning rate is a parameter that determines to what extent a learning system is based on past experiences، and the impact of current events on it. This method is efficient but does not properly function in dynamic and non-stationary environments. In this paper، a new method for adaptive learning rate adjustment in learning automata is proposed. In this method، in addition to the length of time to learn، some statistical characteristics of actions probability vector of Learning Automata are used to determine the increase or decrease of learning rate. Furthermore، unlike existing methods، during the process of learning، both increase and decrease of the learning rate is done and Learning Automata responds effectively to changes in the dynamic random environment. Empirical studies show that the proposed method has more flexibility in compatibility to the non-stationary dynamic environments and get out of local maximum points and the learned values are closer to the true values1.

One of the most important tool in security field are Intrusion Detection Systems. The main purpose of IDS is to monitor suspicious network traffic and generate alerts. These systems are known to generate many alerts. Analyzing these alerts manually by security expert are time consuming، tedious and error prone. From another point of view false positive alerts have huge share of generated alerts. Identifying attack types and generating correct alerts related to attacks is another problem with IDS. In order to overcome mentioned problems alert management systems was introduced. Alert management systems help security experts to manage alerts and produce a high level view of alerts. In this paper a new alert prioritizing algorithm for IDS Alert Management System proposed that uses the fuzzy logic. The proposed algorithm ranks alerts and detects false positive alert1.

This paper aims to introducing a novel method، to find an initial solution for transportation problem which is one of the well-known and useful models in linear optimization. Finding a good initial solution for has a special importance which results in an impressive decrease in the number of iterations to reach the optimal solution in the main algorithm. The proposed method، according to the remaining capacity of each node and calculating an estimation number of next transitions of the node، will reach better answer through making a better choice in each iteration. This method has an acceptable running time، and implementations show the more prospering performance of this method comparing to the well-known algorithm of this subject (Vogel’s method). It should be noted that by increasing the size of the problem (number of the nodes)، performance of the presented method will increase comparing to the Vogel’s1.

This article tries to shortly examine principal component analysis (PCA) in order to extract principal component of an image which would be used for measuring compression rating. Principal Component Analysis is a statistical method which maps high dimension input data to a low dimension output space In PCA a reversible linear transform matrix maps n-dimensional input data to a different dimensional space. Principal Component Analysis can be used to reduce data size. The most important components of data sets that have the greatest impact on the variance will be kept. As mentioned in Rumelhart theory، Feed Forward neural networks can map and compress N orthogonal pattern to logN Pattern in the middle layer. A method called (PCF) has been offered in this article through combining PCA algorithm and Feed Forward Neural Networks in order to compress images. In this method special vectors of image are compressed and the compression rate is developed in comparison with PCA method1.

Clustering is one of the main operations in data mining and its aim is to group similar objects in clusters. This technique seeks to discover structure of dataset by considering similarities or differences between data. Clustering algorithms can be divided into several categories including partitioning clustering algorithms، hierarchical algorithms and density based algorithms. In this paper we investigate some partitioning algorithms and consider them in term of some important parameters and finally compare them1. Keywords — delay/fault-tolerant mobile sensor network، delivery delay، delivery probability، DFT-MSN، erasure coding، pervasive information gathering، queuing theory، replication، transmission overhead