With the growth of online social network platforms and applications, large amounts of textual user-generated content are created daily in the form of comments, reviews, and short-text messages. As a result, users often find it challenging to discover useful information or more on the topic being discussed from such content. Machine learning and natural language processing algorithms are used to analyze the massive amount of textual social media data available online, including topic modeling techniques that have gained popularity in recent years. This paper investigates the topic modeling subject and its common application areas, methods, and tools. Also, we examine and compare five frequently used topic modeling methods, as applied to short textual social data, to show their benefits practically in detecting important topics. These methods are latent semantic analysis, latent Dirichlet allocation, non-negative matrix factorization, random projection, and principal component analysis. Two textual datasets were selected to evaluate the performance of included topic modeling methods based on the topic quality and some standard statistical evaluation metrics, like recall, precision, F -score, and topic coherence. As a result, latent Dirichlet allocation and non-negative matrix factorization methods delivered more meaningful extracted topics and obtained good results. The paper sheds light on some common topic modeling methods in a short-text context and provides direction for researchers who seek to apply these methods.

This paper presents a VLSI implementation of Discrete Wavelet Transform (DWT). The architecture is systolic in nature and performs both h i g h -p a s s and l o w -p a s s coefficient calculations with only one set of multipliers, in contrast to the approaches presented in the literature [ l ] , [2], [3]. The architecture is simple, modular, and cascadable, and has been implemented in VLSI.Experimental results show that real-time coefficient calculation on a 512 X 512 monochrome video input can he achieved with 1.2 pm technology.

A neural network architecture that can be trained to classify e.c.g. is presented. It uses a feature extractor to characterize the e.c.g. before the presentation to a back-propagation network for classification. n e test results indicate that network may accurately classify most sample tapes in AHA database after it has k e n trained on only four sample tapes.

No abstract

Real-time recurrent learning (RTRL), commonly employed for training a fully connected recurrent neural network (RNN), has a drawback of slow convergence rate. In the light of this deficiency, a decision feedback recurrent neural equalizer (DFRNE) using the RTRL requires long training sequences to achieve good performance. In this paper, extended Kalman filter (EKF) algorithms based on the RTRL for the DFRNE are presented in state-space formulation of the system, in particular for complex-valued signal processing. The main features of global EKF and decoupled EKF algorithms are fast convergence and good tracking performance. Through nonlinear channel equalization, performance of the DFRNE with the EKF algorithms is evaluated and compared with that of the DFRNE with the RTRL.

No abstract

This paper presents a VLSI implementation of Discrete Wavelet Transform (DWT). The architecture is systolic in nature and performs both high-pass and low-pass coefficient calculations with only one set of multipliers. The architecture is simple, modular, and cascadable, and has been implemented in VLSI. Simulation results show that real-time coefficient calculation on a 512 X 512 monochrome video'input can be achieved.

Though wavelet transforms have been used to extract bearing fault signatures from vibration signals in the literature, detection results often rely on a proper wavelet function and deep wavelet decomposition. The selection of a proper wavelet function is time consuming and deep decomposition demands more computing effort. This is unsuitable for on-line fault detection. As such, we propose a joint wavelet lifting scheme and independent component analysis (ICA) approach to detecting weak signatures of bearing faults. The optimal envelope spectrum of independent components for signature extraction is selected based on the maximum energy and total energy of each independent component. The performance of the proposed method is evaluated by comparing with several other methods using both simulated and real vibration signals. The results reveal that the proposed method is more effective and robust in extracting bearing fault signatures. The following advantages of the proposed method have also been observed: (a) it is insensitive to wavelet selection and hence is less susceptible to ill selected wavelet function; (b) it is insensitive to the depth of wavelet decomposition, leading to an efficient algorithm; and (c) it takes advantage of ICA in fault detection without using multiple sensors as required in the original ICA.