Phishing attack is now a big threat to people's daily life and networking environment. Through disguising illegal URLs as legitimate ones, attackers can induce users visit the phishing get private information other benefits. Effective methods of detecting websites are urgently needed alleviate threats posed by attacks. As active learning capability from massive data sets, neural network widely used detect However, in stage training many useless small influence features will trap model into...

This paper initiates an innovative concept and basic measurements on testing the IQ (Intelligence Quotient) Internet search engines. It first proposes stipulation of 2014 intelligence scale designs test question bank to aim at engine. To show its applicability, then carries out seven classic engines, such as Google, Baidu, Sogou, Bing, Zhongsou, panguso, so, etc., compared with results same three groups Children whose ages are 6, 12, 18. Based absolute relative ten objects, this finds that...

Biclustering of gene expression data is an important technology for biologists. The biclustering problem proven to be NP-hard. Several methods have been proposed analyze the including genetic algorithms (GAs). However, usually converge slowly when they are used solve largest problem. In this paper, we present a new method, EDA-GA hybrid, data. After testing on simulated data, find hybrid algorithm not only can quickly, but also obtain global solution.

The problem of parallel machine multi-objective scheduling subjected to special process constraint in the textile industries, as one most important combinational optimization problems, is different from other problems following characteristics. On hand, processing machines are non-identical; on sort job processed every can be restricted Considering either minimizing maximum completion time among all machines(makespan) or total earliness/tardiness penalty jobs has been cornerstone studies...

Adaptive blind equalization has gained widespread use in communication systems that operate without training signals. The multi-modulus algorithm can overcome the phase offset error need of any rotator at end equalizer stage while well-known constant modulus cannot. desire for further reduction computational complexity motivated improvement (MMA). This paper presents a modification MMA based on signed-error results more computationally efficient algorithm-signed-error (SE-MMA). performance...

In this paper, a parallel genetic algorithm for finding all roots of complex functional equation based on Parallel Virtual Machine (PVM) is present, research are made in some technical problems realizing. We describe the design and implement equation.

In recent years, the application of multi-objective evolutionary algorithms (MOEAs) to overlapping community detection in complex networks has been a hot research topic. However, existing MOEAs for detecting communities show poor scalability large-scale due fact that encoding length individuals is usually equal number all nodes network. To this end, we suggest reduced mixed representation based algorithm named RMR-MOEA detection, where individual recursively as evolution proceeds....

Nowadays, the demand for wireless spectrum resources is growing rapidly with continuous development of communication technology. However, vast majority available spectrums have been allocated, making new demanders difficulty to acquire their appropriate spectrums. In order reallocate scanty demanders, many auction schemes that are based on truthfulness proposed. insufficient protective measures deployed in these made users private information be easily exposed during process auction. By...

The choice of the optimal clustering number (Kopt) is very important for forming results. As a kind unsupervised learning method, it still difficult to determine Kopt. validity index (CVI) an effective method determining Kopt and evaluating results generated by algorithms. However, many existing CVIs cannot efficiently with overlapping datasets. In this paper, we propose WCH, new cluster WCH constructed based on features inner-cluster tightness, inter-cluster separation can effectively...

Genetic algorithm (GA) is a kind of that simulates the process and mechanism evolution. Because its unique biologic feature suitability to any function, it becomes very popular has been used in many problems fields. Estimation distribution (EDA) an generated from GAs. Comparing with GAs, EDAs replace crossover mutation operations GAs learning sampling probability best individuals population at each iteration algorithm. superior, hot topic recently. Based on former researches, this paper...

Abstract A dual‐mode multi‐modulus algorithm (DM‐MMA) and a stop‐and‐go (SAG‐DMMMA) for blind equalisation of high‐order quadrature amplitude modulation (QAM) signals are proposed. Simulation results show the proposed algorithms have faster convergence speed smaller steady‐state mean square error, compared with recently introduced algorithm. Copyright © 2008 John Wiley & Sons, Ltd.

This paper proposes an improved boost learning algorithm, the SceBoost and its application in developing fast robust features for citrus canker detection by machine vision. The algorithm use symmetric cross entropy to eliminate redundancy among selected using AdaBoost algorithm. Selected are subjected recognize symptoms on given pictures of foliage. Compared with related feature selection our method can get improvements classification accuracy significantly reduce computation time when reach...

This paper investigates how to classify the type of aircraft based on aviation communication signal. Hilbert-Huang Transform (HHT) is introduced adaptively decompose cockpit voice into several IMFs, then IMFs can be used as features for identifying type. To alleviate problem mode mixing caused by signal intermittency in EMD, a novel method EEMD applied here decomposing short-wave communication. The results show that performance HHT better than original EMD decomposition and classification.