International Journal of Modern Physics C

We demonstrate an automatic procedure for extracting features such as directionality of crack patterns, distribution of node distances and segment lengths, fractal dimension, entropy, and crack coverage to aid in automatic classification of painting cracks or craquelures. To test our classifier, we make use of four distinct craquelure patterns, designated by names based on their country of origin, namely: Dutch, Flemish, French or Italian. We report that selecting features based on effect size ratio from the above statistical measures, the standard linear discriminant analysis (LDA) can make predictive classification of the craquelure patterns with a 69.4% accuracy. Effect size ratio simultaneously quantifies the extent of correlation and variance of two statistical sets of data. This test set accuracy is more than two times better than mere chance classification, or the proportional chance criterion, computed to be Φ_PCC = 27.61% and also twice the recommended classifier accuracy 1.25 × Φ_PCC = 34.4%. We compare the result with the nonlinear method of neural network and we observe no marked improvement in the resulting accuracy. This suggests that the problem at hand with respect to the statistical features extracted is a linear classification problem. The work provides a comprehensive guide on the algorithms that can be used to extract quantitative information of crack patterns.

Transmission efficiency and robustness are two important properties of various networks and a number of optimization strategies have been proposed recently. We propose a scheme to enhance the network performance by adding a small fraction of links (or edges) to the currently existing network topology, and we present four edge addition strategies for adding edges efficiently. We aim at minimizing the maximum node betweenness of any node in the network to improve its transmission efficiency, and a number of experiments on both Barabási–Albert (BA) and Erdös–Rényi (ER) networks have confirmed the effectiveness of our four edge addition strategies. Also, we evaluate the effect of some other measure metrics such as average path length, average betweenness, robustness, and degree distribution. Our work is very valuable and helpful for service providers to optimize their network performance by adding a small fraction of edges or to make good network planning on the existing network topology incrementally.

We examined the sector dynamics of Korean stock market in relation to the market volatility. The daily price data of 360 stocks for 5019 trading days (from January, 1990 to August, 2008) in Korean stock market are used. We performed the weighted network analysis and employed four measures: the average, the variance, the intensity, and the coherence of network weights (absolute values of stock return correlations) to investigate the network structure of Korean stock market. We performed regression analysis using the four measures in the seven major industry sectors and the market (seven sectors combined). We found that the average, the intensity, and the coherence of sector (subnetwork) weights increase as market becomes volatile. Except for the "Financials" sector, the variance of sector weights also grows as market volatility increases. Based on the four measures, we can categorize "Financials," "Information Technology" and "Industrials" sectors into one group, and "Materials" and "Consumer Discretionary" sectors into another group. We investigated the distributions of intrasector and intersector weights for each sector and found the differences in "Financials" sector are most distinct.

Cluster analysis is an important way to ascertain whether or not a complex system consists of sub-clusters with different properties. On the basis of journal citation pattern, we clustered 1896 SCI journals in an emerging way (emerging clustering) in this paper. By this emerging clustering, the data to be checked for clustering is reduced from to O(N_J) (N_J is the total number of journals), this reduction could reduce the time complexity of clustering to a certain extent. During the clustering process, characteristic numbers of clusters are obtained, which correspond to various resolution scales for viewing the journals and might be helpful in understanding the mutual interactions among various knowledge domains. Its efficiency is further checked by comparisons between emerging clustering and hierarchical clustering with different inter-cluster linkage techniques. Statistical properties and comparisons of the clustering results show that this is an efficient clustering method for journal system. It could be a helpful improvement for traditional classification methods based on subjective analysis.

The spatial structure has often been identified as a prominent mechanism that substantially promotes the cooperation level in prisoner's dilemma game. In this paper we introduce a weighting mechanism into the spatial prisoner's dilemma game to explore the cooperative behaviors on the square lattice. Here, three types of weight distributions: exponential, power-law and uniform distributions are considered, and the weight is assigned to links between players. Through large-scale numerical simulations we find, compared with the traditional spatial game, that this mechanism can largely enhance the frequency of cooperators. For most ranges of b, we find that the power-law distribution enables the highest promotion of cooperation and the uniform one leads to the lowest enhancement, whereas the exponential one lies often between them. The great improvement of cooperation can be caused by the fact that the distributional link weight yields inhomogeneous interaction strength among individuals, which can facilitate the formation of cooperative clusters to resist the defector's invasion. In addition, the impact of amplitude of the undulation of weight distribution and noise strength on cooperation is also investigated for three kinds of weight distribution. Current researches can aid in the further understanding of evolutionary cooperation in biological and social science.

We have examined the order book characteristics and market impact on the Korean stock index futures market (KOSPI 200 index futures). The distribution of order volumes generally follows power-law distribution. The estimated exponents are 1.9 for market order, 2.5 for limit order, and 2.1 for cancel order. This result is different from the case of stocks where the exponent of market order is larger than that of limit order. The order likelihood is distinctively high in every 50's of order volume, which implies the behavioral characteristics of human preference on round-up numbers. The distributions of bid–ask spread and the best quotes volume provide the evidence of the liquidity of KOSPI 200 index futures market. We have obtained the concave relationship between market impact and transaction volume as well. Finally, the market response behavior is observed regarding various transaction sizes. The size of market response is estimated to be proportional to the size of transaction. Also, the larger the transaction size is, the longer it takes to recover the stability from the impact triggered by transaction.

In this paper, function projective synchronization of chaotic systems is investigated through nonlinear adaptive–impulsive control. To achieve synchronization, suitable nonlinear continuous and impulsive controllers are designed, according to invariant principle of impulsive dynamical systems. Sufficient conditions are given to ensure the synchronization. Numerical simulation results show the effectiveness of the proposed scheme.

This study examines the dependence of crossing time on sequence length for a finite population in an asymmetric multiplicative, or additive, landscape with a positive asymmetric parameter and a fixed extension parameter in three types of mutation-selection model: the coupled discrete-time model, the coupled continuous-time model, and the decoupled continuous-time model. The crossing times for a finite population in the three types of mutation-selection model began to deviate from the crossing times for an infinite population at a critical sequence length. It then increased exponentially as a function of sequence length in the stochastic region where the sequence length was much longer than the critical sequence length. The exponentially increasing rates of the crossing times in the three types of mutation-selection model were similar to each other. These rates were decreased by increasing the asymmetric parameter. Once the asymmetric parameter reached a certain limit the crossing times for the three models in the stochastic region could not be decreased further by increasing the asymmetric parameter past this limit.