Due to the Covid-19 pandemic, most university classes were moved to online instruction. This greatly stimulated the need for online learning tools. WeBWorK is an open source online homework system, which has been used extensively in a variety of subjects. However, it has not been widely adopted by the Computer Science education community. In this paper, we discuss our experience using WeBWorK in teaching two large online sections of discrete mathematics. Emphasis is given to how we created randomized and auto-graded problems for many topics. In addition, we summarize student performance and feedback. We conclude with our reflections on using WeBWorK and propose future work for exploring its adaptive learning features.

5G non-public network (NPN) can provide customized and dedicated network services for various vertical industries. The operation of 5G NPN is a crucial aspect for the deployment and application of 5G NPN. This paper studies the development of 5G NPN operation. Furthermore, this paper proposes a three-stage evolution path, framework and the guaranteed requirements for 5G NPN operation. Some examples are also provided to achieve NPN optimization goal by the framework. The paper provides insights and guidance for the vertical industries of 5G NPN operation, as well as suggests potential directions for future work on 5G NPN operation.

The growth analysis of enterprises is an important basis for predicting the future development trend of enterprises. For an enterprise itself, the enterprise growth analysis can help the enterprise to understand its own business situation. It can also assist the enterprise to accurately customize the development strategy. As far as the investment market is concerned, the enterprise growth analysis can help investors comprehensively understand the investment target and reduce the investment risk as well as improve the investment benefit. This paper makes a comparative analysis on the growth of 4937 enterprises with all A-shares in different industries from seven dimensions, including competitiveness, profitability, operation ability, debt paying ability, R & D ability, scale expansion ability, enterprise supply chain ability. This paper reveals that there are significant differences in the growth of enterprises in different industries in the post-epidemic era.

Recently, information and digital technology are widely used in thousands of industries, leading to intelligent transformation, traditional methods, which lacks intelligent instrument. The safety of college students has attracted widespread attention from all walks of life, while campus safety management still adopts manual and traditional methods, which lacks intelligent instrument and big data resources and technologies are not fully utilized. In this paper, we propose a system of campus safety management based on telecom big data and data fusion architecture, providing solutions for intelligent campus management. In addition, a prediction algorithm of student behavior intent considering time spans has been proposed, proving the advantages in accuracy metrics and F1-score compared with traditional prediction algorithms.

This paper mines user behavior characteristics based on big data technology. This paper proposes a method for behavior slicing based on historical activity data, and insights into the personalized behavior characteristics. Firstly, the data is processed, and the classification is expanded on the basis of the parsed APP label types. Secondly, a time slicing method is proposed to reduce information loss, which integrates time, location, business type, and behavior into individual users. Then, based on slices of a day and a week, the paper analyzes user behavior and construct a portrait of user’s interest and preference. Finally, the periodic factor method is utilized to predict the behavior changes, forming the feature labels for users. Based on real business behaviors, this paper provides insight into user personality and effectively improves the authenticity and accuracy of prediction.

Recently, location-based service has become a hot research topic. Mobile communication data records abundant information about users’ temporal and spatial characteristics. By modeling the users’ mobility based on mobile communication data, this can assist to understand human user patterns more accurately and deeply. Initially, this paper introduces three mainstream algorithms for user mobility modeling. Then this paper proposes a novel Markov chain based user mobility prediction scheme. The proposed scheme is implemented through four stages, including time and space division, Markov property examination, transition probability matrix calculation, Markov model weighting. Experimental results show that the proposed scheme can achieve higher accuracy compared with the traditional algorithms.

Enterprises loss is a growth strategy, in which enterprises migrate across regions/cities to adapt to the changes of internal and external environment, in this way to seek new development space and further reach the growth again. As the carrier of local economic development, the transfer of enterprises from one region to another undoubtedly means the loss of regional resources for the region. This paper takes large- scale enterprises as the research object. Then, this paper uses questionnaire data and statistical data, and adopts the combination of PCA algorithm and extreme value standardization method to comprehensively evaluate the loss probability of enterprises. This method will reflect the loss tendency of enterprises in the region, and make an empirical analysis on the large-scale enterprises in region, in this way to help regional managers have an early insight into the loss tendency of enterprises in the region. Finally, it will provide a reference for stabilizing the regional economy and help reduce the loss risk of large-scale enterprises in the region.

We have established a research-practice partnership (RPP) to build a computer science (CS) and computational thinking (CT)-focused STEM ecosystem at two middle schools. Creating such an ecosystem to broaden student participation in computing through an RPP approach involves all stakeholders in the research process. Borrowing upon visual participatory research methods, we developed a graphic research instrument to engage teachers in the research process and elicit their perspectives on strategies for building the ecosystem. This experience report describes our research methodology across two distinct cases to demonstrate the utility of this drawing activity as an investigative and partnership development tool. The contribution is in offering a flexible approach to other university-based RPP teams that enables a synergistic partnership development tool and data collection instrument that can be tailored to a variety of RPP contexts, facilitating more productive and equitable ways of engaging stakeholders in the research process. We describe our project contexts and share results from the pilot study with practitioner-members of our RPP teams. We discuss two cases to highlight the contribution this approach made to the development of our partnerships.

Social network marketing is a very promising topic in the data operation work of telecom operators. Based on the big data collection and analysis of telecom operators, this paper presents a content recommendation scheme which considering both users' social relationships and users' personal preferences. Regarding users' personal preference analysis, this scheme uses DPI (Deep Packet Inspection) technology to obtain the user's personal preference tag and evaluate the user's preference index. In terms of user social relations, it integrates the analysis of mobile users' communication behaviors, temporal and spatial relationships, interaction circles and other related indicators. Logistic regression algorithm is used to illustrate the influence from a user to another. The preference recommendation scheme based on the mobile network user social circle proposed in this paper expands the value scenarios of operators' big data, integrates resources and channels, improves operators' data insight capabilities, and realizes the value mining and enhancement of operators' big data.

Psychological health of students has become a widespread social problem, while the management and assessment of college students' psychological health is still stay in passive and manual mode based on the traditional method. In this paper, we design a novel architecture for the prediction of college students' psychological health based on Multi-Source big data including Operation Support System big data, educational data and psychological health questionnaire data. Then we propose the Optimized Decision Tree using Multiple-Target Particle Swarm Optimization (DT-MTPSO) algorithm. Experiment shows that the proposed algorithm can solve the Multiple-Target problems effectively and has better performance in F1-score than traditional Decision Tree. In addition, the result of the features selection of DT-MTPSO for different targets shows the relationship between the psychological health level and behavioural characteristics of students for different evaluation indicators, providing guidance to the school managers and educational psychologist.

With the rapid development of Internet industry, people are facing increasing challenge of information overload. Under this background, personalized recommendation has been comprehensively researched in order to provide a more time-saving and accurate way for information retrieval. In this paper, a novel hybrid recommendation scheme based on collaborative filtering and association rules is put forward to compensate the weaknesses of individual algorithms. This scheme is implemented through several steps. Firstly, it solves the problem of data sparsity with the help to association rules, and then employs the revised collaborative filtering to calculate the similarity among the items. Finally, it predicts user ratings for the unknown items based on item similarity and generates recommendation lists according to the prediction ratings. Experimental results show that the recommendation accuracy of this hybrid scheme has been dramatically improved compared to other traditional algorithms.

This paper studies the Internet characteristics of Internet media news based on three factors, including journalism, communication, statistical physics. By obtaining the indicators (e.g., news release time, title, text, media, media type, media weight, etc.), the news transmission speed evaluation system is constructed by employing clustering model, and the news transmission speed is further obtained. Through this indicator and system, it can effectively reflect the transmission speed and changing trend of news events, and monitor the spread situation of news events in real time. In addition, this paper provides a reference basis for the governance of network public opinion and the early warning as well as handling for crisis events.

This project is a Research to Practice Partnership (RPP) between two middle schools and two universities. It focuses on investigating problems and on identifying solutions around increasing participation and interest in computer science (CS). We aim to do this by identifying, experimenting with, and fine-tuning methods to help students develop computational thinking (CT) skills. The research employs a STEM ecosystem model, which facilitates a support structure that aims to mitigate barriers and impact students as they progress in STEM areas. While this RPP is still a work in progress, we present data from the first year of our collaboration with one of the middle schools. While the research questions guiding this RPP are intended to be iterative and revised annually, year one data provides perspectives on (1) barriers to developing a STEM ecosystem that supports CS/CT for every student through integration into science, math, and language arts courses, (2) the factors or interventions needed for the development of a CS/CT focused ecosystem that supports everyone in the school, (3) the indicators of success for a CS/CT focused STEM ecosystem in a school, and (4) how the ecosystem prepares and engages all students for CS/CT work in high school. Year one data is discussed in terms of the STEM ecosystem framework and in how it will guide the next steps in this partnership. This project contributes to the understanding of how to prepare future generations for participation in a workforce where knowledge of the foundations of CS/CT is integral to success.

In recent years, many cutting-edge research results have emerged in the field of computer vision, especially in the field of image classification. But, researchers are still trying to explore more sophisticated models to further improve the accuracy of image classification. However, many models have failed to achieve satisfactory results due to the complexity of the image and the problems of image noise. Therefore, in order to solve these problems, this paper proposes an unsupervised feature extraction method called multi-granularity convolution denoising autoencoder (MGCDAE). Based on convolutional neural network, the method proposed the concept of multi-granularity convolution kernel to solve the problem of complex image feature extraction. In addition, we introduce denoising autoencoder (DAE) for image noise, which enables our approach to extract more robust features from noise images. The high-level features extracted by the above method are sent to the softmax classifier for classification to evaluate the validity of the feature extraction. Our method has been evaluated on three benchmark data sets, which results show that our approach can extract more discriminative high-level features compared with other existing algorithms.

In mobile crowdsensing (MCS), to select the optimal set of participants for a particular sensing task, the cloud-based MCS platform requires mobile users to submit their bids and their sensing quality data. This can cause privacy breaches. One possible solution is to leverage secure sharing or bidding schemes to protect participants' personal information during selection. However, these schemes suffer from high overheads, poor scalability and more importantly, the group formation has never been studied. To address this issue and to enhance the protection of user privacy, we propose a set of novel privacy-preserving grouping methods, which place participants into small groups over hierarchical edge clouds. By doing this, not only can the participants be hidden in groups, but also the overall privacy-preserving participant selection becomes more scalable. The design goal to minimize the communication cost during secure sharing/bidding within groups, while satisfying each participant's requirement for privacy preservation. For different scenarios and optimization functions, we propose a set of grouping schemes to fulfill this goal. Extensive simulations over both synthetic and real-life datasets illustrate the efficiency of proposed mechanisms.

It has been long established that discrete structures is an important and foundational component of the computer science curriculum. However, the topics covered in this course tend to be more abstract than those covered in most other introductory computer science courses. This leads to additional challenges for instructors and students. To deal with these challenges, we introduce a new pedagogy for teaching this course. Our approach is based on a variant of the flipped classroom paradigm and is comprised of four main components: before class preparatory work, in-class mini lecture, in-class team based problem solving activities, and weekly individual assignments. In this paper, we discuss these components in detail. Our approach is informed by several cutting-edge teaching methodologies including active learning, light weight teams, and gamification. We conclude the paper by discussing the results of a survey taken by the students and a summary of the grades attained in the class. These show that our approach was well received by the students and has led to good learning outcomes.

Auction based participant selection has been widely used for mobile crowd sensing (MCS) to achieve user incentive and assignment optimization. However, mobile crowd sensing problems solved with auction-based approaches usually involve participants' privacy concerns because a participant's bids may contain her private information (such as location visiting patterns), and disclosure participants' bids may disclose their private information as well. In this paper, we study how to protect such bid privacy in a temporally and spatially dynamic MCS system. We assume that both sensing tasks and mobile participants have dynamic characteristics over spatial and temporal domains. Following the classical VCG auction, we carefully design a scalable grouping based privacy-preserving participant selection scheme, which leverages Lagrange polynomial interpolation to perturb participants' bids within groups. The proposed solution does not affect the operation of current MCS platform. Both theoretical analysis and real-life tracing data simulations verify the efficiency and security of the proposed solution.

Mobility models are crucial for the simulation and evaluation of protocols for multihop wireless networks. However, most commonly used mobility models do not reflect the way humans actually move. This significantly affects the reliability of simulation results. In this paper we introduce a new mobility model called the Smoothly Truncated Levy Walk (STLW), which is more realistic than most standard models that appear in the literature. Its main innovations are as follows. First, to take into account dependencies in the direction of motion, it models changes in the direction instead of the standard approach, which directly models the direction and ignores these dependencies. Second, it uses realistic models for pause times, flight lengths, and changes in direction. In particular, it uses tempered stable distributions to model pause times and flight lengths and the beta distribution to model changes in direction. We justify the use of these distributions from both a theoretical and an empirical perspective. In particular, we perform a trace-based validation on several real-world traces from various scenarios. Validation results show that this model is very flexible and can be used to model human movements in a variety of situations.

Applications for wireless sensor networks have grown enormously over the past few years. Routing and sink discovery protocols designed for ad hoc networks do not adapt to sensor networks, and generic sensor network routing techniques are not optimal for all scenarios. We propose a light weight routing protocol to discover mobile sinks in battle-field operations or large terrains. Initial experimentation shows promising results in reduction of control over head, good data delivery with support of multi-metric path selection.

The wireless multihop to an access point model appears to be a promising component of future network architectures, including multihop cellular networks and wireless access networks at the edges of mesh networks. Sophisticated software radios and core network protocols are being developed to support the integration of heterogeneous air interfaces within these access networks. A key challenge is managing diverse resources at access points (e.g., 3G, WiFi or WiMax) while discovering efficient multi-hop paths from a source to an access point based on selection criteria specified by various applications or necessitated by network resource constraints. We propose a new routing protocol that integrates multiple metrics to calculate path cost based on diverse selection criteria. In addition, a hybrid proactive/reactive anycast routing paradigm is applied to guide the discovery of an access point among multiple available access points. The result is an integrated, flexible protocol for route discovery and access point discovery. Simulation analysis shows that our approach outperforms single-metric routing protocols while supporting flexible service criteria, including load balancing at access points.

Topology control protocol aims to efficiently adjust the network topology of wireless networks in a self-adaptive fashion to improve the performance and scalability of networks. This is especially essential to large-scale multihop wireless networks (e.g., wireless sensor networks). Fault-tolerant topology control has been studied recently. In order to achieve both sparseness (i.e., the number of links is linear with the number of nodes) and fault tolerance (i.e., can survive certain level of node/link failures), different geometric topologies were proposed and used as the underlying network topologies for wireless networks. However, most of the existing topology control algorithms can only be applied to two-dimensional (2D) networks where all nodes are distributed in a 2D plane. In practice, wireless networks may be deployed in three-dimensional (3D) space, such as under water wireless sensor networks in ocean or mobile ad hoc networks among space shuttles in space. This article seeks to investigate self-organizing fault-tolerant topology control protocols for large-scale 3D wireless networks. Our new protocols not only guarantee k -connectivity of the network, but also ensure the bounded node degree and constant power stretch factor even under k −1 node failures. All of our proposed protocols are localized algorithms, which only use one-hop neighbor information and constant messages with small time complexity. Thus, it is easy to update the topology efficiently and self-adaptively for large-scale dynamic networks. Our simulation confirms our theoretical proofs for all proposed 3D topologies.

The calculation of path cost is a critical component of route discovery for network routing. The criteria used to represent path cost guides resource consumption in the network. In this paper, we describe our approach, a set of protocols based on our Multiple Metrics Routing Protocol (MMRP) for integrating hop count, energy consumption, and traffic load into the path cost calculation for ad hoc or multihop-cellular networks. Our initial aim is to select among multiple disjoint routes in order to maintain a low path cost, in terms of energy consumption and delay, without depleting resources at popular intermediate nodes. One extension of MMRP removes the constraint that only disjoint paths are considered and enables discovery of more optimal routes. A second extension includes adaptive adjustment of cost metrics to support device classification (e.g., energy capacity, bandwidth) in heterogeneous networks. We illustrate our approach with a simple example, followed by extensive simulation analysis. Results indicate that proper combination of multiple metrics for calculating path costs results in improved performance and lower overall system resource consumption as compared to AODV or energy efficient routing protocols.

The wireless multihop to an access point model appears to be a promising component of future network architectures, including multihop cellular networks and wireless access networks at the edges of mesh networks. Sophisticated software radios and core network protocols are being developed to support the integration of heterogeneous access networks. A challenge is managing diverse resources at access points (e.g., 3G, WiFi or WiMax), as well as the distributed interference among the mobiles within a heterogeneous access network. We propose the use of a new anycasting protocol to guide access point discovery and path selection for balancing access point resource and routing packets to access points (APs). In addition, a hybrid proactive/reactive approach is used to reduce overhead of AP discovery. We use theoretical analysis and extensive simulations, to study the tradeoff of our hybrid anycasting protocol. The simulation study indicates that the use of the hybrid anycasting protocol for AP discovery, load balancing and routing, results in consistent performance improvements.

Fault tolerant topology control in wireless networks has been studied recently. In order to achieve both sparseness (i.e., the number of links is linear with the number of nodes) and fault tolerance (i.e., can survive certain level of node/link failures), different geometric topologies were proposed and used as the underlying network topologies for wireless networks. However, most of the existing topology control algorithms can only be applied to 2-dimensional (2D) networks where all nodes are distributed in a 2D plane. In practice, wireless networks may be deployed in 3-dimensional (3D) space, such as underwater wireless sensor networks in ocean or ad hoc networks in space. This paper seeks to investigate efficient fault tolerant topology control protocols for 3D wireless networks. Our new protocols not only guarantee the kappa-connectivity of the network, but also ensure the bounded node degree and constant power stretch factor. All of our proposed protocols are localized algorithms, which only use one-hop neighbor information and constant messages with small time complexity. Our simulation confirms our theoretical proofs for all proposed 3D topologies.

Energy aware routing protocols are consistently cited as efficient solutions for ad hoc and sensor networks routing and data management. However, there is not a consistent approach to define the energy related cost metrics that are used to guide the routing protocol performance. This paper provides a survey and analysis of energy related metrics used for ad hoc routing. First, the most common energy efficient routing protocols are classified into four categories based on the energy cost metrics employed. Then, the results of our simulation-based analysis are presented. We conducted a complete set of simulations to compare and contrast the performance of various energy-related metrics. Our analysis provides a comparison of the performance of energy cost metrics used within AODV-based ad hoc routing protocols.

Future network architectures for mesh networks, sensor networks, and hybrid networks include multiple wireless hops with fixed or mobile routers relaying messages between mobile nodes and access points. The vast literature on routing for ad hoc networks is applicable for route discovery between a mobile node and an access point for these multi-hop networks. However, most ad hoc routing protocols rely upon "least-cost" or "shortest-path" routing, wherein a single metric of hop-count or energy-usage is used to define the path cost. For multi-hop networks, a number of additional metrics should be simultaneously considered for determining path cost. In this paper, we introduce our first step towards combining a number of cost criteria to define path cost during route discovery. Our simulation results indicate that giving simultaneous consideration to hop-count, path-congestion, and energy-usage results in a system performance with greater packet delivery ratio, lower end-to-end delay, lower overhead, and lower per-node energy usage

Sequential minimal optimization (SMO) is one popular algorithm for training support vector machine (SVM), but it still requires a large amount of computation time for solving large size problems. This paper proposes one parallel implementation of SMO for training SVM. The parallel SMO is developed using message passing interface (MPI). Specifically, the parallel SMO first partitions the entire training data set into smaller subsets and then simultaneously runs multiple CPU processors to deal with each of the partitioned data sets. Experiments show that there is great speedup on the adult data set and the Mixing National Institute of Standard and Technology (MNIST) data set when many processors are used. There are also satisfactory results on the Web data set.

This paper proposes a modified version of support vector machines (SVMs), called c-ascending support vector machines (c-ASVMs), to model non-stationary financial time series. c-ASVMS are obtained by a simple modification of the regularized risk function in SVMs whereby the recent /spl epsiv/-insensitive errors are penalized more heavily than the distant /spl epsiv/-insensitive errors. This procedure is based on the prior knowledge that in the non-stationary financial time series, the recent past data could provide more important information than the distant past data. In the experiment, c-ASVMS are tested using three real futures collected from the Chicago Mercantile Market. It is shown that the c-ASVMS with the actually ordered sample data consistently forecast better than the standard SVMs, with the worst performance when the reversely ordered sample data are used. Furthermore, the c-ASVMs use fewer support vectors than those of the standard SVMs, resulting in a sparser representation of solution.

Recently, support vector machine (SVM) has become a popular tool in time series forecasting. In developing a successful SVM forecaster, the first important step is feature extraction. This paper applies kernel principal component analysis (KPCA) to SVM for feature extraction. KPCA is a nonlinear PCA developed by using the kernel method. It firstly transforms the original inputs into a high dimensional feature space and then calculates PCA in the high dimensional feature space. By examining the sunspot data and one real futures contract, the experiment shows that SVM by feature forms much better than that extraction using KPCA per without feature extraction. In comparison with PCA, there is also superior performance in KPCA.

A novel type of learning machine called support vector machine (SVM) has been receiving increasing interest in areas ranging from its original application in pattern recognition to other applications such as regression estimation due to its remarkable generalization performance. This paper deals with the application of SVM in financial time series forecasting. The feasibility of applying SVM in financial forecasting is first examined by comparing it with the multilayer back-propagation (BP) neural network and the regularized radial basis function (RBF) neural network. The variability in performance of SVM with respect to the free parameters is investigated experimentally. Adaptive parameters are then proposed by incorporating the nonstationarity of financial time series into SVM. Five real futures contracts collated from the Chicago Mercantile Market are used as the data sets. The simulation shows that among the three methods, SVM outperforms the BP neural network in financial forecasting, and there are comparable generalization performance between SVM and the regularized RBF neural network. Furthermore, the free parameters of SVM have a great effect on the generalization performance. SVM with adaptive parameters can both achieve higher generalization performance and use fewer support vectors than the standard SVM in financial forecasting.