Abstract The Internet has recently become not only one of the most popular communication channels but also the most accessible and searchable information repository of different domains. Billions of users surf the Internet everyday to search for information or visit social network and e-commerce websites. The web usage behaviors of these users can be analyzed by Web Usage Mining (WUM) systems... The Internet has recently become not only one of the most popular communication channels but also the most accessible and searchable information repository of different domains. Billions of users surf the Internet everyday to search for information or visit social network and e-commerce websites. The web usage behaviors of these users can be analyzed by Web Usage Mining (WUM) systems to discover useful knowledge that helps improve service performance. Despite recent successes, existing WUM systems still cannot cope with the growing dynamics and complexity of the Web, resulting in overwhelming overheads and low efficiency. In this paper, an innovative Web-Page Recommender System is proposed to model user web browsing behaviors, extract popular web paths and predict web navigation possibilities. Particularly, the main inference algorithm in this system integrates advanced Frequent Pattern Recognition methods and Stochastic Markov Models to achieve an optimal balance between superior predictive accuracy and excessively demanding computation in higher order models. Empirical analysis suggests that our system outperforms other conventional methods with respect to complexity reduction and accuracy improvement. Read More Read Less

Abstract Many issues exist in the communities based on Product Data Management. To address and solve the issues, many applications are being developed and applied. One importnat and quite useful application develops from the Artificial Intelligence. In this research paper we address the importance of Product Data Management (PDM) with respect to the industrial contributional point of view. Moreover we also... Many issues exist in the communities based on Product Data Management. To address and solve the issues, many applications are being developed and applied. One importnat and quite useful application develops from the Artificial Intelligence. In this research paper we address the importance of Product Data Management (PDM) with respect to the industrial contributional point of view. Moreover we also present some currently available major challenges to the Product Data Management based communities, and targeting those challenges we discuss an already proposed artificial intelligence based approach .i.e., I-SOAS, and briefly describe how this approach can be helpful in solving the PDM communities faced problems. Read More Read Less

Abstract Currently, the recommender systems are found to be beneficial in many applications with an increase of combination of other technologies. One such technology is the soical networks where Collaborative Filtering (CF) is effective in providing collaborative social environment. The major issue is building the reliabel statistics for estimating trust among actors involved in the social networks. In this work, we... Currently, the recommender systems are found to be beneficial in many applications with an increase of combination of other technologies. One such technology is the soical networks where Collaborative Filtering (CF) is effective in providing collaborative social environment. The major issue is building the reliabel statistics for estimating trust among actors involved in the social networks. In this work, we propose a design for creating an effective collaborative filtering on social network and a new approach for estimating trust factors. This approch emerges by using the data extracted from The Facebook Project I. The statistics trust data have the potential as better input parameters in Collaborative Filtering algorithm. We have also used the extracted data as the case study to document the proposed system and parameters. Read More Read Less

Abstract Datamining has potential applications in several fields including the Business Intelligence (BI) where in the datamining is useful in for understanding the trends and reacting to events. We advocate that the crucial area in dataming, the stream-mining where continuous data streams arrive into the system and get mined very quickly. It further induces the framework for creating a new type... Datamining has potential applications in several fields including the Business Intelligence (BI) where in the datamining is useful in for understanding the trends and reacting to events. We advocate that the crucial area in dataming, the stream-mining where continuous data streams arrive into the system and get mined very quickly. It further induces the framework for creating a new type of real-time Business Intelligence architecture. The algorithmic level stream mining and digital information system architecture was addressed and applied separately. We now try to focus a single view on the real-time Business Intelligence system architecture powered by stream-mining. Besides we present a few more possible applications in which the proposed architecture can support are discussed.. Read More Read Less

Abstract The volume of Web content is increasing and hence the heterogeneity also grows. The structure denies the understanding the content which leads to difficulty in obtaining effective results for the end users. Web research now becomes directed to get high relevancy in web searching. One important technique is the introductiopn of the interactive graphical and visualization techniques which enable to... The volume of Web content is increasing and hence the heterogeneity also grows. The structure denies the understanding the content which leads to difficulty in obtaining effective results for the end users. Web research now becomes directed to get high relevancy in web searching. One important technique is the introductiopn of the interactive graphical and visualization techniques which enable to handle large numbers of results while simultaneously presenting several attributes for each Web page. Besides the query reformulation and reconstruction is usually controlled by the search engine. In many web search processes, the twin problems are the redundancy and/or irrelevancy. To ensure the effectiveness in the overall search relevance the user is integrated in the process of query reformulation. This study has used the Visual Search Engine (the VSE) where the query reformulation and results presentation are visualized. In the user study, the effectiveness of the VSE was demonstrated when compared to Google. Read More Read Less