STRUCT

Set intersection is an important operation and widely used in both database and graph analytics applications. However, existing state-of-the-art set intersection methods only consider the size of input sets and fail to optimize for the case in which the intersection size is small. In real-world scenarios, the size of most intersections is usually orders of magnitude smaller than the size of the input sets, e.g., keyword search in databases and common neighbor search in graph analytics. In this paper, we present FESIA, a new set intersection approach on modern CPUs. The time complexity of our approach is O(n/√w + r), in which w is the SIMD width, and n and r are the size of input sets and intersection size, respectively. The key idea behind FESIA is that it first uses bitmaps to filter out unmatched elements from the input sets, and then selects suitable specialized kernels (i.e., small function blocks) at runtime to compute the final intersection on each pair of bitmap segments. In addition, all data structures in FESIA are designed to take advantage of SIMD instructions provided by vector ISAs with various SIMD widths, including SSE, AVX, and the latest AVX512. Our experiments on both real-world and synthetic datasets show that our intersection method achieves more than an order of magnitude better performance than conventional scalar implementations, and up to 4x better performance than state-of-the-art SIMD implementations.

Keyword search in relational databases has recently emerged as a new research topic. As a search result is often assembled from multiple relational tables, existing IR-style ranking strategies can not be applied directly. In this paper, we propose a novel IR ranking strategy considering query semantics for effective keyword search. The experimental results on a large-scale real database demonstrate that our method results in significant improvement in terms of retrieval effectiveness as compared to previous ranking strategies.

As an active research field, database keyword search (KWS) has put much emphasis on the performance issues, due to its high computational cost. However, a closer examination on KWS reveals other interesting aspects worth noting. In this paper, we examine KWS from a broader perspective, particularly from its relationship with data mining. Freed from syntax-related considerations, KWS users now have better opportunities to explore the data in the way as they wish, and such exploration may reveal useful hindsight for what to be done in data mining. Recently we have conducted our KWS research from this unique perspective. We propose a software environment which offers a dual-mode approach to explore KWS: the database mode allows the implementation of database KWS directly by incorporating various KWS algorithms, while the XML mode converts the database contents to an XML document on which KWS is conducted. The dual mode approach not only has the potential of achieving integrated KWS on both structured and semistructured data, but also facilitates query relaxation by incorporating ontologies in the XML mode. The software environment still allows us to observe performance related issues of KWS; but more importantly, it offers a freehand approach for users to explore the data, thus has the potential of aiding data mining. Component design and experimental studies are described.

Keyword search in relational databases allows the user to search information without knowing database schema and using structural query language. As results needed by user are assembled from connected tuples of multiple relations, ranking keyword queries are needed to retrieve relevant results. For a given keyword query, the authors first generate candidate networks and also produce connected tuple trees according to the generated candidate networks by reducing the size of intermediate joining results. They then model the generated connected tuple trees as a document and evaluate score for each document to estimate its relevance. Finally, the authors retrieve top-k keyword queries by ranking the results. In this paper, the authors propose a new ranking method based on virtual document. They also propose Top-k CTT algorithm by using the frequency threshold value. The experimental results are shown by comparison of the proposed ranking method and the previous ranking methods on IMDB and DBLP datasets.

Thematic ranking of object summaries for keyword search

Aims and objectives. The aim of this literature review was to examine empirical research to date on hospital discharge and illuminate areas in need of further exploration. Background. Discharging older people from hospital has long been associated with difficulties and complications. With a steady increase in the ageing population in Ireland and their use of acute hospital services, the successful management of hospital discharge is now a central concern to health service management. Methods. Using a seven-step approach by Stolz et al. [Scandinavian Journal of Caring Sciences (2004) Vol. 18, 111-119] a keyword database search was conducted which revealed 44 studies that matched the author's inclusion criteria. Results. These studies uncover aspects of the discharge process that are in need of further research, and particularly suggest more rigour in the measurement of hospital discharge outcomes. Relevance to clinical practice. This review focuses on a very important clinical issue given the predicted increase in the use of acute services by older people and the known vulnerabilities of the interface between hospital and community settings.

In this work, we derive the semantics of a given English query and convert it into its equivalent SQL query. Instead of using neural networks for semantic analysis of English queries, we opt for a Context Free Grammar approach. Most neural-network-based systems can handle only one semantic at a time, whereas, because of the flexibility offered by our CFG approach, our system manages to handle simultaneous usage of conjunctive, disjunctive, and negative semantics. It also handles complex statements comprising of main as well as dependent clauses. In addition, the system also takes into account aggregate functions and constructs the required GROUP-BY and HAVING clauses. We describe how the system analyzes English queries by understanding the role that each part-of-speech has to play in constructing SQL queries. Numerous examples demonstrate the effectiveness of our approach where state-of-the-art techniques relying on deep learning algorithms fail to deliver.

Keyword-based search in relational database is an easy and effective way for ordinary users or Web users to access relational databases. Even though relational database management systems (RDBMs) have provided full-text search capabilities, they do not support keyword-based search model. The text databases and relational databases are different that is a challenging task to apply the keyword search techniques in information retrieval (IR) to DB. A common method to performing keyword search in relational database is to generate the minimum connected tuple sets in schema graph transformed from relations. Although existing candidate network (CN) generation methods retrieve a set of joining tuples, they are still problem which is causing large overhead for CNs generation. In this paper, we propose a new candidate network generation algorithm (Heuristic_CNGen) based on the iterative deepening A* (IDA*) algorithm. The proposed algorithm produces a minimum number of CNs according to the maximum number of tuple set. We generate CNs for a given keyword query. And then, we identify the connected tuple tree as a result according to generated CNs. We evaluate the proposed method on DBLP.

Social media networks contain both content and context-specific information. Most existing methods work with either of the two for the purpose of multimedia mining and retrieval. In reality, both content and context information are rich sources of information for mining, and the full power of mining and processing algorithms can be realized only with the use of a combination of the two. This paper proposes a new algorithm which mines both context and content links in social media networks to discover the underlying latent semantic space. This mapping of the multimedia objects into latent feature vectors enables the use of any off-the-shelf multimedia retrieval algorithms. Compared to the state-of-the-art latent methods in multimedia analysis, this algorithm effectively solves the problem of sparse context links by mining the geometric structure underlying the content links between multimedia objects. Specifically for multimedia annotation, we show that an effective algorithm can be developed to directly construct annotation models by simultaneously leveraging both context and content information based on latent structure between correlated semantic concepts. We conduct experiments on the Flickr data set, which contains user tags linked with images. We illustrate the advantages of our approach over the state-of-the-art multimedia retrieval techniques.

Technical troubleshooting is a domain that has changed enormously in recent years. Instead of relying on visits from service personnel end users facing technical problems with machinery, for example computers and printers, can now seek assistance from systems that guide them toward an autonomous solution of the problem. Systems that can be offered to them are wide in their range, but typically fall either in the category of Expert Systems or searchable databases that can be queried with keyword searches. Both approaches present advantages and disadvantages in terms of flexibility to address different levels of user expertise and ease of maintenance. However, few studies explicitly address the issue of how best to design for a balance between guidance and user freedom in such systems. In the work reported here an office equipment manufacturer’s call centre was studied in order to understand the mechanisms used when human agents guide users toward a resolution. The overall aim here is not to reproduce the agent behaviour in a system, but rather to identify which interactional building blocks such a system should have. These are assessed in relation to the existing online knowledge base resources offered by the same company in order to exemplify the kinds of issues designers need to attend to in this domain.

The goal of the RENDEZVOUS Project is to enable a casual user to access relational data bases in natural language without a preliminary training period. The system will, if necessary, engage in an English dialogue with the user in order to arrive at an unambiguous statement of his query. This dialogue serves the triple purpose of aiding the user in his problem specification, of clarifying ambiguities inherent in the English language and of enabling the system to overcome incomprehensible features of the users query. The conversation must constantly narrow the scope of inquiry, resorting to menu selection if necessary until the user and system come to an agreement as to what information has been requested. At any moment during the dialogue, the system must be capable of expressing its current understanding of the user's query in clear and unambiguous English. In particular, the system must be able to rephrase the user's query for approval before execution.

CD-CARS: Cross-domain context-aware recommender systems

It has become desirable to provide a way of keyword search for users to query structured information in an XML database (data-centric retrieval) by combining database and information retrieval techniques. Therefore, the key challenges of keyword search in the XML database are how to define appropriate result models meeting user’s search intents, how to search the results by using efficient algorithms, and how to ranking the results. In this chapter, on one hand, the authors present the foundational knowledge of XML keyword search such as XML data models, XML query languages, inverted index, and Dewey encoding. On the other hand, some existing typical researches of keyword search in XML are presented, including the results models such as Smallest Lowest Common Ancestor (SLCA), Exclusive Lowest Common Ancestor (ELCA), Meaningful Lowest Common Ancestor (MLCA), the related search algorithms, and the ranking approaches.

The aim of this study is to analyse the development of publications of scientific articles focusing on the project-based learning model. The method used is a literature review with a bibliometric analysis approach. A keyword search in the Scopus journal database for “Project-Based Learning; Science Education” was used to start the research. An analysis was carried out using the Vosviewer application, which allows in-depth exploration of trends and key contributions in the field. The results showed that a total of 702 publications were recorded in Scopus-indexed journals over a fifteen-year period, from 2010 to 2024. The United States emerged as the country that contributed the most publications, followed by Indonesia in second place, reflecting its active role in promoting research related to the Project based learning model. Among researchers, Capraro, Robert M. was identified as the most prolific contributor during the period 2010-2024. In addition, the keyword ‘project-based learning’ appeared most frequently in the articles, confirming its centrality in discussions of innovative educational practices. These findings suggest a significant and growing interest in Project Based Learning in science education. However, this study highlights the need for a broader scope of research. Future research should not be limited to Scopus-indexed journals, but should also include various databases and resources.

Purpose: The purpose of this paper is to review and present that separate capabilities of knowledge management (KM) are identified and classified as sequential capability. These KM capabilities are utilized to affect organizational performance with a KM-cyclic view. Design/methodology/approach: This paper is a literature review research, through which KM capability and performance are identified, synthesized, from related books, literature and other research studies. Using key-word search in electronic databases, we generated a fairly exhaustive list of the articles dealing with the topic of KM capability in the period of 1979 to 2010 as well as various sources were extended from searching results. A total of 60 articles were identified for KM capability and 30 articles for performance measurement. Findings: The results of the research suggest two main aspects: 1) The KM capabilities were organized as four categories with cyclic view: acquisition/creation, conversion, sharing, and application. 2) The KM performance measurements were classified as two categories: non-financial and financial measurements. Research limitations/implications: There is a need to undertake empirical research and in-depth case studies of knowledge management practices using a KM cyclic view as a framework with which to interpret findings. Originality/value: The KM cyclic view is used to make some recommendations on appropriate research approaches to further our understanding of relationship of KM capability and performance in research. This also provides future researchers with useful means to assess the KM performance in different KM capability and to realize the benefits of a knowledge-based economy.