In association rule mining, frequent itemset mining (FIM) optimization is moving from software to hardware acceleration. High utility itemset mining (HUIM), which is an advanced FIM extension, solves traditional FIM's inability to handle highvalue data well. However, the hardware acceleration of HUIM is more complex and memory-intensive, demanding stricter FPGAbased acceleration design, leaving FPGA-accelerated HUIM research preliminary. To address this, this paper proposes an FPGA coprocessor for HUIM acceleration. It includes a multiplexed computation unit that enhances join operation efficiency and is scalable to parallel modules, and a prefix-free itemset method that reduces redundant computation and access via hardware traits. Experimental results demonstrate that the proposed CPUFPGA heterogeneous parallel processing architecture achieves a speedup of up to$5 \times$compared to pure software-based HUIM methods

ABSTRACT Driven by the rise of the internet, recent years have witnessed the gradual manifestation of commercial values of online reviews. In movie industry, sentiment analysis serves as the foundation for mining user preferences among diverse and multi‐layered audiences, providing insight into the market value of movies. As a representative task, aspect‐based sentiment analysis (ABSA) aims to analyse and extract fine‐grained sentiment elements and their relations in terms of discussed aspects. Relevant studies, particularly in the realm of deep learning research, face challenges due to insufficient annotated data. To alleviate this problem, we propose a guideline for fine‐grained sentiment annotations that defines aspect categories, describes the method for annotating aspect sentiment triplets, either simple or complex and designs a scheme to represent hierarchical labels. Based on this, an ABSA dataset tailored for the movie domain is curated by annotating on 1100 Chinese short reviews acquired from Douban. Applicability of both the annotation guideline and curated data is evaluated through inter‐annotator consistency and self‐consistency checks, and domain adaptation assessment of e‐commerce and healthcare cases. Predictive performance of machine learning models on this dataset shed light on possible applications in more fine‐grained sentiment analysis in the movie domain, for example, figuring out the aspects from which to stimulate viewership and influence public opinions, thereby providing substantial support for the movie's box office performance. Finally, we extended our fine‐grained sentiment annotation guideline to the e‐commerce and healthcare. Through empirical experimentation, we demonstrated the universality of these guideline across diverse domains.

Network function virtualization (NFV) in Elastic Optical Inter-DataCenter Networks (EO-DCNs) enables a flexible, adaptive, effective, and economic network services deployment and upgrade. However, it is facing critical threats from large-scale network failures due to natural disasters. This is driving the need for efficient network protection schemes of service function chain (SFC) provisioning. In this paper, we investigate the disaster-resilient SFC provisioning problem leveraging power-efficient path protection with distance-adaptive modulation format (MF) assignment, concerning virtual network function (VNF) placement, SFC mapping, path protection, constrained recovery delay, and spectrum allocation simultaneously. An integer linear program (ILP) model is formulated to jointly minimize power consumption and spectrum usage, subject to disaster resilience. A heuristic algorithm is further developed for the sake of scalability. Numerical simulation results demonstrate that the proposed disaster protection schemes enable saving up to 32.05% power consumption.

Network function virtualization (NFV) in Elastic Optical Inter-DataCenter Networks (EO-DCNs) enables a flex-ible, adaptive, effective, and economic network services de-ployment and upgrade. However, it is facing critical threats from large-scale network failures, due to natural disasters. This is driving the need for efficient network protection schemes of service function chain (SFC) provisioning. In this paper, we investigate the disaster-resilient SFC provisioning problem leveraging power-efficient path protection with distance-adaptive modulation format (MF) assignment, concerning virtual network function (VNF) placement, SFC mapping, path protection, and spectrum allocation simultaneously. An integer linear program (ILP) model is formulated to jointly minimize power consumption and spectrum usage, subject to disaster resilience. A heuristic algorithm is also developed for the sake of scalability. Numerical Simulation results demonstrate that the proposed disaster pro-tection schemes enable saving up to 32.05 % power consumption.

As the increasing traffic patterns show asymmetric feature, directed pre-configured-cycle (p-cycle) has indicated the ability of better protection in elastic optical networks (EONs). In this paper, we investigate three different integer linear program (ILP) models of directed p-cycle without candidate cycle enumeration leveraging flow conservation. Three directed p-cycle designs are based on the same directed p-cycle strategy but differ from each other in how the flows can construct the directed p-cycles, namely individual link flow (ILF) directed p-cycle, aggregated link flows (ALF) directed p-cycle, and loop-eliminating flow (LEF) directed p-cycle, respectively. These ILPs aim to jointly minimize power consumption and spectrum usage of all directed p-cycles configured in EONs. The problem formulation involves directed p-cycle generation, modulation format (MF) selection, power consumption optimization, and spectrum allocation. Furthermore, the proposed directed p-cycle strategy is designed with a compact and novel MF adaptation relying on accurate protection path lengths. Simulations are conducted to compare the proposed ILPs with the conventional method which uses a rough upper bound on MF adaptation. Numerical results demonstrate that all of the three proposed ILPs have better performances on the joint objective, in which the improvement is up to 24.31%. Although the proposed ILPs are with the same performance on the objective due to the same directed p-cycle strategy, the LEF directed p-cycle shows the best efficiency.

Natural disasters have challenged the survivability of Elastic Optical Inter-DataCenter Networks (EO-DCNs), and it is urgent to establish efficient disaster protection schemes. In this paper, we investigate the disaster-resilient service provisioning problem leveraging cooperative storage system (CSS). Instead of mirrored content backup on a single DC, our proposed CSS partitions a required content into no less than three fragments if possible, each of which is then stored on a DC located in different disaster zones. Accordingly, multi-path routing with the adaptive number of working paths to distinct DCs is employed to serve each request, while a protection path is computed to protect against a disaster failure. Our main objective is to jointly minimize the spectrum usage and maximal occupied frequency slot index (MOFI) subject to disaster resilience. Besides, we also expect to cut the content storage space. To this end, we propose for the first time a CSS-based dedicated end-to-content path protection (CDP), which allows service provisioning through multiple paths with the adaptive number of paths rather than a single path. This consequently reduces at least half of the reserved spectrum on the protection path. To find the optimal CDP strategy, we formulate the studied problem as an integer linear program (ILP) and then propose a fast heuristic algorithm. Observing the trade-off between the spectrum usage and content storage space, we further design a maximum-CDP (M-CDP), which generates the maximum number of working paths to reduce the content storage space. Simulations are conducted to compare the proposed schemes with the traditional protection strategy using mirrored storage and single-path routing. Numerical results demonstrate that the proposed CSS-based protection schemes enable to cut up to 21.6% of the spectrum usage and 15% of the content storage space.

Natural disasters have challenged the survivability of Elastic Optical Inter-DataCenter Networks (EO-DCNs), and it is urgent to establish efficient disaster protection schemes. In this paper, we investigate the disaster-resilient service provisioning problem leveraging cooperative storage system (CSS) and multipath routing. The studied problem involves data center (DC) assignment, content partition and placement, working/protection paths computation, as well as spectrum allocation. Our main objective is to jointly minimize the spectrum usage and maximal frequency slot index. Besides, we also expect to cut the content storage space. To this end, we first formulate the studied CSS-based protection problem as an integer linear program (ILP), and then propose a fast heuristic algorithm to improve the network scalability in large instances. Numerical simulations are conducted to compare the proposed schemes with the traditional protection strategy using entire content replication and single path routing. Simulation results demonstrate that the CSS-based protection scheme enables to cut up to 17.8% of the spectrum usage and half of the content storage space.

In this paper, we propose a novel network selection method oriented to users' quality-of-experience (QoE) for indoor visible light communication (VLC) heterogeneous networks. The proposed method considers the difference between user requirements for different parameters and the actual performance of each candidate network in these reference indicators. To improve QoE, a new indicator named “benefit-cost-ratio (BCR)” is defined to represent the user demand under different businesses and assist in sorting alternatives to select the optimal network for access. Simulation results show that through the proposed network selection scheme, the candidate networks could be effectively ranked and adjusted according to user requirements.