Vol. 39 · No. 5

Feature selection plays a vital role to reduce the high dimension of the feature space in the text document classification problem. The dimension reduction of feature space reduces the computation cost and improves the text classification system accuracy. Hence, the identification of a proper subset of the significant features of the text corpus is needed to classify the data in less computational time with higher accuracy. In this proposed research, a novel feature selection method which combines the document frequency and the term frequency (FS-DFTF) is used to measure the significance of a term. The optimal feature subset which is selected by our proposed work is evaluated using Naive Bayes and Support Vector Machine classifier with various popular benchmark text corpus datasets. The experimental outcome confirms that the proposed method has a better classification accuracy when compared with other feature selection techniques.

In this paper, we discuss the application of convolutional neural networks (CNNs) for hierarchical text classification using local top-down approaches. We present experimental results implementing a local classification per node approach, a local classification per parent node approach, and a local classification per level approach. A 20Newsgroup hierarchical training dataset with more than 20 categories and three hierarchical levels was used to train the models. The experiments involved several variations of hyperparameters settings such as batch size, embedding size, and number of available examples from the training dataset, including two variation of CNN model text embedding such as static (stat) and random (rand). The results demonstrated that our proposed use of CNNs outperformed flat CNN baseline model and both the flat and hierarchical support vector machine (SVM) and logistic regression (LR) baseline models. In particular, hierarchical text classification with CNN-stat models using local per parent node and local per level approaches achieved compelling results and outperformed the former and latter state-of-the-art models. However, using CNN with local per node approach for hierarchical text classification underperformed and achieved worse results. Furthermore, we performed a detailed comparison between the proposed hierarchical local approaches with CNNs. The results indicated that the hierarchical local classification per level approach using the CNN model with static text embedding achieved the best results, surpassing the flat SVM and LR baseline models by 7

Spiking neural P systems (SNP systems, in short) are a class of distributed parallel computation systems, inspired from the way that the neurons process and communicate information by means of spikes. A new variant of SNP systems, which works in asynchronous mode, asynchronous spiking neural P systems with multiple channels and symbols (ASNP-MCS systems, in short), is investigated in this paper. There are two interesting features in ASNP-MCS systems: multiple channels and multiple symbols. That is, every neuron has more than one synaptic channels to connect its subsequent neurons, and every neuron can deal with more than one type of spikes. The variant works in asynchronous mode: in every step, each neuron can be free to fire or not when its rules can be applied. The computational completeness of ASNP-MCS systems is investigated. It is proved that ASNP-MCS systems as number generating and accepting devices are Turing universal. Moreover, we obtain a small universal function computing device that is an ASNP-MCS system with 67 neurons. Specially, a new idea that can solve ``block'' problems is proposed in INPUT modules.

MapReduce is a programming paradigm for large-scale distributed information processing. This paper proposes a MapReduce algorithm for the minimum vertex cover problem, which is known to be NP-hard. The MapReduce algorithm can efficiently obtain a minimal vertex cover in a small number of rounds. We show the effectiveness of the algorithm through experimental evaluation and comparison with exact and approximate algorithms which demonstrates a high quality in a small number of MapReduce rounds. We also confirm from experimentation that the algorithm has good scalability allowing high-quality solutions under restricted computation times due to increased graph size. Moreover, we extend our algorithm to randomized one to obtain a good expected approximate ratio.

Hyperspectral image (HSI) classification has been a hot topic in the remote sensing community. A large number of methods have been proposed for HSI classification. However, most of them are based on the extraction of spectral feature, which leads to information loss. Moreover, they rarely consider the correlation among the spectrums. In this paper, we see spectral information as a sequential data which should be relevant to each other. We introduce long short-term memory (LSTM) model, which is a typical recurrent neural network (RNN), to deal with HSI classification. To tackle the problem of overfitting caused by limited labeled samples, regularization strategy is introduced. For unbalance in different classes, we improve LSTM by weighted cost function. Also, we employ guided filter to smooth the HSI that can greatly improve the classification accuracy. And we proposed a method for modeling hyperspectral sequential data, which is very useful for future research work. Finally, the experimental results show that our proposed method can improve the classification performance as compared to other methods in three popular hyperspectral datasets.

Software projects are almost always team efforts and successful projects involve well-formed and well-composed teams. Past studies have revealed that personality contributes to effective team composition and, therefore, project success. Yet despite its importance, only a couple of empirical studies have quantitatively evaluated the impact of personality on software quality and team productivity. Our previous study was an effort in this direction. In that study, we proposed a metric called Team Homogeneity Index and evaluated its impact on software quality and team productivity for two phases (implementation and testing) of the software development life cycle. This study is a continuation of our previous work. In this study, we replicate our experiment on three different phases of software development life cycle (i.e. analysis and design, implementation, and testing). We also determine the weights for all five personality traits using input from the industry and propose an improved version of Team Homogeneity Index called Weighted Team Homogeneity Index. Finally, we conduct a comparative analysis of Team Homogeneity Index and Weighted Team Homogeneity Index to determine whether weights assigned to personality traits make any difference. Our findings reveal that weights do make a difference and Weighted Team Homogeneity Index is more strongly correlated than Team Homogeneity Index for almost all of the teams, especially those composed of practitioners, in the three different phases of Software Development Life Cycle.

For critical embedded systems the formal validation and verification is required. However, the real-time model checking suffers from problems of state-space explosion and clock explosion. The aim of this paper is to ensure an improvement of the Modeling and Analysis of Real-Time Embedded systems (MARTE), which is de facto standard, with formal semantics for verification finality. Therefore, we propose an operational method for translating UML sequence diagrams with MARTE annotations to Time Petri nets with Action Duration specifications (DTPN). Based on true concurrency semantics, the semantics of these specifications are defined in terms of Duration Action Timed Automata (daTA).

For the exact detection of moving targets in video processing, an adaptive low-rank sparse decomposition algorithm is proposed in this paper. In the paper's algorithm, the background model and the solved frame vector are first used to construct an augmented matrix, then robust principal component analysis (RPCA) is used to perform a low-rank sparse decomposition on the enhanced augmented matrix. The separated low-rank part and sparse noise correspond to the background and motion foreground of the video frame, respectively, the incremental singular value decomposition method and the current background vector are used to update the background model. The experimental results show that the algorithm can deal with complex scenes such as light changes and background motion better, and the algorithm's delay and memory consumption can be reduced effectively.

As the rapid development of online transactions, transaction frauds have also emerged seriously. The fraud strategies are characterized by specialization, industrialization, concealment and scenes. Anti-fraud technologies face many challenges under the trend of new situations. In this paper, aiming at sample imbalance and strong concealment of online transactions, we enhance the original deep forest framework to propose a deep forest-based online transaction fraud detection model. Based on the BaggingBalance method we propose, we establish a global sample imbalance processing mechanism to deal with the problem of sample imbalance. In addition, the autoencoder model is introduced into the detection model to enhance the representation learning ability. Via the three-month real online transactions data of a China's bank, the experimental results show that, evaluating by the metric of precision and recall rate, the proposed model has a beyond 10

Energy dissipation is a prominent factor for the very large scale integrated circuit (VLSI). The reversible logic-based circuit was capable to compute the logic without energy dissipation. Accordingly, reversible circuits are an emerging domain of research based on the low value of energy dissipation. At nano-level design, the critical factor in the logic computing paradigm is the fault. The proposed methodology of fault coverage is powerful for testability. In this article, we target three factors such as fault tolerance, fault coverage and fault detection in the reversible KMD Gates. Our analysis provides good evidence that the minimum test vector covers the 100