Short text clustering is one of the fundamental tasks in natural language processing. Different from traditional documents, short texts are ambiguous and sparse due to their short form and the lack of recurrence in word usage from one text to another, making it very challenging to apply conventional machine learning algorithms directly. In this article, we propose two novel approaches for short texts clustering: collapsed Gibbs sampling infinite generalized Dirichlet multinomial mixture model infinite GSGDMM) and collapsed Gibbs sampling infinite Beta-Liouville multinomial mixture model (infinite GSBLMM). We adopt two flexible and practical priors to the multinomial distribution where in the first one the generalized Dirichlet distribution is integrated, while the second one is based on the Beta-Liouville distribution. We evaluate the proposed approaches on two famous benchmark datasets, namely, Google News and Tweet. The experimental results demonstrate the effectiveness of our models compared to basic approaches that use Dirichlet priors. We further propose to improve the performance of our methods with an online clustering procedure. We also evaluate the performance of our methods for the outlier detection task, in which we achieve accurate results. 相似文献
Pattern Analysis and Applications - A bounded multivariate generalized Gaussian mixture model with a full covariance matrix is proposed for modeling data in a bounded support region. For model... 相似文献
Virtual Reality - This study aims to fill a gap in current research on virtual reality (VR) by developing a valid and reliable educational VR acceptance scale based on the unified theory of... 相似文献
Virtual Reality - The aim of the study is to address a gap in the literature by developing an educational virtual reality (edVR) attitude measurement instrument, which determines college... 相似文献
The condensation of naphthalene-1- or 2-carbaldehyde with dimethyl 2,2-dimethyl-succinate (Stobbe conditions) gives predominantly the ( Z )-hemiesters 3 and 11a . The hemiester 3 contains a small amount of the ( E )-isomer. Similar condensation of naphthalene-2-carbaldehyde with dimethyl-2-methylsuccinate gave the hemiester ( E )- 12a . Their configuration and the ratios of the ( Z:E )-hemiesters, their derived acids, anhydrides, indenones and phenanthrene derivatives are inferred by high resolution 1H-n.m.r. spectroscopy. 相似文献
Neural Computing and Applications - As a result of various loads, including critical installations (industries, nuclear facilities, etc.), electrical distribution networks (EDNs) must operate... 相似文献
Using Fusarium oxysporum species (F. oxysporum), a green synthesis of super-paramagnet iron oxide @silver@ Chitosan (SPION@Ag@Cs) was achieved. The physico-chemical characteristics of the SPION@Ag@Cs nanocomposite revealed the development of superparamagnetic phases of iron oxide (Fe2O3), as well as fluorescence, Raman absorption, and biocompatibility. Drug and gene delivery, and diagnosis, are all possibilities for the nanocomposite. The uptake of nitric oxide by HT-29 colorectal cell lines is examined in this article, for up to 72 h, the cytotoxicity of the HCT116 cell line was investigated. These characteristics were compared to Streptomyces griseus fungal species (S. griseus) which develops Fe3O4 under the same preparation conditions.
Cloud Computing (CC) is the most promising and advanced technology to store data and offer online services in an effective manner. When such fast evolving technologies are used in the protection of computer-based systems from cyberattacks, it brings several advantages compared to conventional data protection methods. Some of the computer-based systems that effectively protect the data include Cyber-Physical Systems (CPS), Internet of Things (IoT), mobile devices, desktop and laptop computer, and critical systems. Malicious software (malware) is nothing but a type of software that targets the computer-based systems so as to launch cyber-attacks and threaten the integrity, secrecy, and accessibility of the information. The current study focuses on design of Optimal Bottleneck driven Deep Belief Network-enabled Cybersecurity Malware Classification (OBDDBN-CMC) model. The presented OBDDBN-CMC model intends to recognize and classify the malware that exists in IoT-based cloud platform. To attain this, Z-score data normalization is utilized to scale the data into a uniform format. In addition, BDDBN model is also exploited for recognition and categorization of malware. To effectually fine-tune the hyperparameters related to BDDBN model, Grasshopper Optimization Algorithm (GOA) is applied. This scenario enhances the classification results and also shows the novelty of current study. The experimental analysis was conducted upon OBDDBN-CMC model for validation and the results confirmed the enhanced performance of OBDDBN-CMC model over recent approaches. 相似文献