Abstract The video game industry has grown rapidly, driven by increased online gaming engagement and diverse game genres. To help users discover potential games amid data overload, effective recommendation models are essential for online platforms. Although recent research addresses cold start issues by integrating multimodal contents, current approaches focus on the implicit modeling of collaborative connections between items and users, neglecting latent semantic associations within content attributes. Additionally, they fail to effectively integrate diverse modalities, relying on simplistic fusion techniques that overlook intricate content. This paper proposed a novel generative AI-augmented multimodal graph-based recommender with parallel-stage contrastive learning framework (GAIMG-CL), constructs the latent semantic item-item and user-user relationships based on image and textual data and incorporates the concept of multimodal contrastive learning. By leveraging modality-aware latent relationships, graph convolutions are employed to integrate both item and user correlations into explicit modality-aware representations. Moreover, our innovative contrastive learning framework enhances fine-grained multimodal fusion. We also leverage GAI for valuable information extraction, enhancing representation learning quality. This approach also enables the integration of enhanced item and user representations into collaborative filtering techniques. Our proposed approach outperforms other recommendation methods and notably boosts cold start recommendation performance, as evidenced by the experimental results.

Abstract With air pollution having become a global concern, scientists are committed to working on its amelioration. In the field of air pollution prediction, there have been good results in experimental research so far, but few studies have integrated weather forecast information and the properties of air pollution drift. In this work, we propose a novel wind‐sensitive attention mechanism with a long short‐term memory (LSTM) neural network model to predict the air pollution ‐ PM2.5 concentrations by considering the influence of wind direction and speed on the changes of spatial–temporal PM2.5 concentrations in neighbouring areas. Preliminary predictions for PM2.5 are then made by an LSTM neural network regarding neighbouring pollution; these predictions are “paid attention to” and we finally apply an ensemble learning method based on eXtreme Gradient Boosting (XGBoost) to combine the preliminary predictions with weather forecasting to make second phase predictions of PM2.5. The experiment is conducted using PM2.5 data and weather forecast data. Our results illustrate that the proposed method is superior to other methods in predicting PM2.5 concentrations, including multi‐layer perceptron, support vector regression, LSTM neural network, and extreme gradient boosting algorithm.

Abstract In virtual worlds (VWs), users have more VW games alternatives, whereas VW companies consequently suffer from high customer turnover rate and low customer loyalty. Therefore, building a churn prediction model to facilitate subsequent churn management and customer retention is important. The churn behaviours and the impact of social neighbour influences to customer churn may be different for different types of users. Accordingly, we segment users into stable, unstable, and solitary user groups according to their social contact behaviours in VWs. Novel segmentation‐based churn prediction approaches are proposed for churn prediction in VWs by building prediction models for each type of user groups and considering the effect of social neighbour influences for different user groups. The proposed approaches are evaluated by conducting experiments with a dataset collected from a VW platform. The experimental results show different churn prediction performances under different user groups. The segmentation‐based churn prediction approaches perform better than do general approaches without considering user groups. Moreover, the results also reveal that social neighbour influences have a positive impact on stable and unstable users. The proposed work contributes to investigating the social neighbour influences on churn prediction for different types of user groups in VWs.

With the ubiquity of the Internet and the rapid development of Web 2.0 technology, social question and answering (SQA) websites have become popular knowledge‐sharing platforms. As the number of posted questions and answers (QAs) continues to increase rapidly, the massive amount of question‐answer knowledge is causing information overload. The problem is compounded by the growing number of redundant QAs. SQA websites such as Yahoo! Answers are open platforms where users can freely ask or answer questions. Users also may wish to learn more about the information provided in an answer so they can use related keywords in the answer to search for extended, complementary information. In this article, we propose a novel approach to identify complementary QAs (CQAs) of a target QA. We define two types of complementarity: partial complementarity and extended complementarity. First, we utilize a classification‐based approach to predict complementary relationships between QAs based on three measures: question similarity, answer novelty, and answer correlation. Then we construct a CQA network based on the derived complementary relationships. In addition, we introduce a CQA network analysis technique that searches the QA network to find direct and indirect CQAs of the target QA. The results of experiments conducted on the data collected from Yahoo! Answers Taiwan show that the proposed approach can more effectively identify CQAs than can the conventional similarity‐based method. Case and user study results also validate the helpfulness and the effectiveness of our approach.

Abstract Mobile data communications have evolved as the number of third generation (3G) subscribers has increased. The evolution has triggered an increase in the use of mobile devices, such as mobile phones, to conduct mobile commerce and mobile shopping on the mobile web. There are fewer products to browse on the mobile web; hence, one‐to‐one marketing with product recommendations is important. Typical collaborative filtering (CF) recommendation systems make recommendations to potential customers based on the purchase behaviour of customers with similar preferences. However, this method may suffer from the so‐called sparsity problem, which means there may not be sufficient similar users because the user‐item rating matrix is sparse. In mobile shopping environments, the features of users' mobile phones provide different functionalities for using mobile services; thus, the features may be used to identify users with similar purchase behaviour. In this paper, we propose a mobile phone feature (MPF)‐based hybrid method to resolve the sparsity issue of the typical CF method in mobile environments. We use the features of mobile phones to identify users' characteristics and then cluster users into groups with similar interests. The hybrid method combines the MPF‐based method and a preference‐based method that uses association rule mining to extract recommendation rules from user groups and make recommendations. Our experiment results show that the proposed hybrid method performs better than other recommendation methods.

Recommender systems can mitigate the information overload problem and help workers retrieve knowledge based on their preferences. In a knowledge‐intensive environment, knowledge workers need to access task‐related codified knowledge (documents) to perform tasks. A worker's document referencing behavior can be modeled as a knowledge flow (KF) to represent the evolution of his or her information needs over time. Document recommendation methods can proactively support knowledge workers in the performance of tasks by recommending appropriate documents to meet their information needs. However, most traditional recommendation methods do not consider workers’ KFs or the information needs of the majority of a group of workers with similar KFs. A group's needs may partially reflect the needs of an individual worker that cannot be inferred from his or her past referencing behavior. In other words, the group's knowledge complements that of the individual worker. Thus, we leverage the group perspective to complement the personal perspective by using hybrid approaches, which combine the KF‐based group recommendation method (KFGR) with traditional personalized‐recommendation methods. The proposed hybrid methods achieve a trade‐off between the group‐based and personalized methods by exploiting the strengths of both. The results of our experiment show that the proposed methods can enhance the quality of recommendations made by traditional methods.

In data mining problems, data is usually provided in the form of data tables. To represent knowledge discovered from data tables, a decision logic (DL) is proposed in rough set theory. DL is an instance of propositional logic, but we can use other logical formalisms to describe data tables. In this paper, we propose two descriptions of data tables based on first-order data logic (FODL) and attribute value-sorted logic (AVSL) respectively. In the context of FODL, we show that explicit definability and implicit definability in classical logic implies the notion of definability in rough set theory. We also show that AVSL is particularly useful for the representation of properties of many-valued data tables.

Effective patent management is essential for organizations to maintain their competitive advantage. The classification of patents is a critical part of patent management and industrial analysis. This study proposes a hybrid-patent-classification approach that combines a novel patent-network-based classification method with three conventional classification methods to analyze query patents and predict their classes. The novel patent network contains various types of nodes that represent different features extracted from patent documents. The nodes are connected based on the relationship metrics derived from the patent metadata. The proposed classification method predicts a query patent's class by analyzing all reachable nodes in the patent network and calculating their relevance to the query patent. It then classifies the query patent with a modified k-nearest neighbor classifier. To further improve the approach, we combine it with content-based, citation-based, and metadata-based classification methods to develop a hybrid-classification approach. We evaluate the performance of the hybrid approach on a test dataset of patent documents obtained from the U.S. Patent and Trademark Office, and compare its performance with that of the three conventional methods. The results demonstrate that the proposed patent-network-based approach yields more accurate class predictions than the patent network-based approach.

The new generation of web-based communities, Web2.0, represents an innovative spirit in sharing and managing contents. Social bookmarking is a portal for users to share, organize, search, and manage bookmarks of web resources. However, with the rapid growth of web documents that are produced every day, people are facing the problem of information overload. The Social bookmarking web site provides the push (user recommendation) counts of articles indicating the recommended popularity degrees of articles. In this paper, we propose to derive the popularity degree of an article by considering the reputation of users that push the article. Moreover, we propose a personalized blog article recommendation approach, which combines the reputation-based popularity with content based filtering, to recommend popular blog articles to users that satisfy their personal preferences. Our experimental results show that the proposed approach outperforms conventional approaches.

Attribute reduction is very important in rough set-based data analysis (RSDA) because it can be used to simplify the induced decision rules without reducing the classification accuracy. The notion of reduct plays a key role in rough set-based attribute reduction. In rough set theory, a reduct is generally defined as a minimal subset of attributes that can classify the same domain of objects as unambiguously as the original set of attributes. Nevertheless, from a relational perspective, RSDA relies on a kind of dependency constraint. That is, the relationship between the class labels of a pair of objects depends on the componentwise comparison of their condition attributes. The larger the number of condition attributes compared, the greater the probability that the constraint will hold. Thus, elimination of condition attributes may cause more object pairs to violate the constraint. Based on this observation, a reduct can be defined alternatively as a minimal subset of attributes that does not increase the number of objects violating the constraint. While the alternative definition coincides with the original one in ordinary RSDA, it is more easily generalized to cases of fuzzy RSDA and relational data analysis.

Abstract For projects in knowledge‐intensive domains, it is crucially important that knowledge management systems are able to track and infer workers' up‐to‐date information needs so that task‐relevant information can be delivered in a timely manner. To put a worker's dynamic information needs into perspective, we propose a topic variation inspection model to facilitate the application of an implicit relevance feedback (IRF) algorithm and collaborative filtering in user modeling. The model analyzes variations in a worker's task‐needs for a topic (i.e., personal topic needs) over time, monitors changes in the topics of collaborative actors, and then adjusts the worker's profile accordingly. We conducted a number of experiments to evaluate the efficacy of the model in terms of precision, recall, and F‐measure. The results suggest that the proposed collaborative topic variation inspection approach can substantially improve the performance of a basic profiling method adapted from the classical RF algorithm. It can also improve the accuracy of other methods when a worker's information needs are vague or evolving, i.e., when there is a high degree of variation in the worker's topic‐needs. Our findings have implications for the design of an effective collaborative information filtering and retrieval model, which is crucial for reusing an organization's knowledge assets effectively.

Rough set theory is an effective tool for data mining. According to the theory, a concept is definable if it can be written as a Boolean combination of equivalence classes induced from classification attributes. On the other hand, definability in logic has been explicated by Beth's theorem. In this paper, we propose two data representation formalisms, called first-order data logic (FODL) and attribute value-sorted logic (AVSL), respectively. Based on these logics, we explore the relationship between logical definability and rough set definability.

In rough set theory, objects are partitioned into equivalence classes based on their attribute values, which are essentially functional information associated with the objects. Therefore, rough set theory can be viewed as a theory of functional granulation. In contrast, relational information systems (RIS) specify the relationships between objects, instead of the properties of objects. In this paper, we present a theory of granulation based on hybrid information systems (HIS), which combine functional information systems (FIS) and RIS. We study the relationship between FIS and RIS. We also define the indiscernibility relation based on relational information, and use it to develop a theory of granulation based on HIS.

Diverse requirements of the participants involved in a business process bring forth the need for a flexible process model that is capable of providing appropriate process information for the various participants. However, the current activity-based approach is inadequate for providing the different participants with varied process information. This paper describes a novel process-view model for workflow management. A process view is an abstracted process derived from a base process to provide abstracted process information. The underlying concepts and a formal model of a process view are presented. Moreover, a novel ordering-preserved approach is proposed to derive a process view from a base process. The proposed approach enhances the flexibility and functionality of conventional activity-based workflow management systems.

Linear algebra-based techniques have long been used to correlate similar documents. They map the documents to a multidimensional vector space, in which each document is represented by a vector. Searching related documents then translates into searching nearest neighbors in the vector space. We propose an indexing structure, called cosine R-tree, which indexes multidimensional vector space and provides efficient nearest neighbor search. Our preliminary results show that it gives better performance than a brute-force linear scan strategy.

Recent advances in video database technology and increasing use of the World Wide Web allow users to access video documents on the Web. Moreover, advances in digital library technologies necessitate the storing of large quantities of video documents. Consequently, video documents must be categorized to assist the user in browsing and searching them by category. This study presents a novel approach to classifying video documents under classes in a predefined class hierarchy. Video documents are classified according to the structural weight of video contents, capturing their characteristics with respect to the hierarchical structure of the video data. Furthermore, while considering the hierarchical relationship among classes, the proposed technique strives for an appropriate balance between specificity and exhaustivity in video classification.

Current spatial database management systems (SDBMS) provide efficient access methods and operators for point and range queries over collections of spatial points, line segments, and polygons. However, it is not clear if existing spatial access methods can efficiently support network computations which traverse line segments in a spatial network based on connectivity rather than geographic proximity. The expected I/O cost for many network operations can be reduced by maximizing the Weighted Connectivity Residue Ratio (WCRR), i.e., the chance that a pair of connected nodes that are more likely to be accessed together are allocated to a common page of the file. CCAM is an access method for general networks that uses connectivity clustering. CCAM supports the operations of insert, delete, create, and find as well as the new operations, get-A-successor and get-successors, which retrieve one or all successors of a node to facilitate aggregate computations on networks. The nodes of the network are assigned to disk pages via a graph partitioning approach to maximize the WCRR. CCAM includes methods for static clustering, as well as dynamic incremental reclustering, to maintain high WCRR in the face of updates, without incurring high overheads. We also describe possible modifications to improve the WCRR that can be achieved by existing spatial access methods. Experiments with network computations on the Minneapolis road map show that CCAM outperforms existing access methods, even though the proposed modifications also substantially improve the performance of existing spatial access methods.

We propose a new similarity-based technique for declustering data. The proposed method can adapt to available information about query distributions, data distributions, data sizes and partition-size constraints. The method is based on max-cut partitioning of a similarity graph defined over the given set of data, under constraints on the partition sizes. It maximizes the chances that a pair of data-items that are to be accessed together by queries are allocated to distinct disks. We show that the proposed method can achieve optimal speed-up for a query-set, if there exists any other declustering method which will achieve the optimal speed-up. Experiments in parallelizing grid files show that the proposed method outperforms mapping-function-based methods for interesting query distributions as well for non-uniform data distributions.>

CCAM is an access method for general networks. It uses connectivity clustering. The nodes of the network are assigned to disk pages via the graph partitioning approach to maximize the CRR, i.e., the chances that a pair of connected nodes are allocated to a common page of the file. CCAM supports the operations of insert, delete, create, and find as well as the new operations, get-A-successor and get-successors, which retrieve one or all successors of a node to facilitate aggregate computations on networks. CCAM includes methods for static clustering, as well as dynamic incremental reclustering, to maintain, high CRR, in the face of updates without incurring high overheads. Experimental analysis indicates that CCAM can outperform many other access methods for network operations.>

Single-fault, multi-fault, 0-1 static sensitizable path, and robust path delay fault are often used to measure the testability of a circuit. We explore the relationships among these testabilities. In addition to the relationships discovered before, we prove that 100% single fault testability, 100% 0-1 static sensitizability are equivalent in two-level single-output circuits. We also prove that 100% 0-1 static sensitizability implies 100% multi-fault testability, and that 100% robust path delay fault testability implies 100% multi-fault testability in two-level circuits. Several new conditions for gate merging while keeping 100% single-fault testability are presented. We further prove that the three transformations D/sub 1,1,2/ (J. Rajski and J. Vasudevamurthy, 1992), extraction, De-Morgan keeping 100% single fault testability also preserve 100% multiple-fault testability, 100% multi-fault testability, 100% robust path delay fault testability. We answer the following two open questions: does 100% multi-fault testability in a multiple outputs circuit not require 100% 0-1 static sensitizability? Does 100% multi-fault testability in a single output circuit imply 100% 0-1 static sensitizability?.>