Catálogo de publicaciones - libros

The implementation of information systems in organizations is a long standing research topic that has preoccupied researchers ever since computer-based information systems started being used in businesses in the early 1960s. However, despite more than 40 years of research, the implementation phenomenon continues to present both practical and research difficulties. The present paper presents a broad overview and assessment of past and current research on implementation, and provides suggestions for future research that will help address some of the challenges implementation researchers currently face.

The communication and information revolution is caused by a fast changing sets of technologies that have already caused changes in the enterprise. Furthermore the expectations of the "customers" of the enterprise have changed because of personal use of the internet and web. Customers expect the Time of response for any transaction to be instantaneous. Managing the pace of change is today’s big enterprise problem. The technologies of computers, networks, software that enable very fast response are complex and themselves fast changing .To use the new versions of technologies requires both learning new skills and changing internal operational procedures. Operational Transformation is the next frontier of business advantage. Because of global competition in uncertain times, Any enterprise of any size must be configured to change; change the way they conduct business and change basic operations. Failure to do this will mean losing to competitors who do change. These issues will be examined and a possible solution to the problem offered.

Enterprise ontology, as defined in this paper, is the essential knowledge of the construction and the operation of the organization of an enterprise, completely independent of the ways in which they are realized and implemented. A methodology is presented for producing the ontology of an enterprise. It has a sound theoretical foundation, which is summarized to the extent needed for understanding the methodology. A small example enterprise serves to illustrate how the methodology works, and to demonstrate the practical usefulness of the notion of enterprise ontology. The motivation for the reported research is the apparent need to make societal institutions, both companies and government agencies, transparent to its customers, as well as to its employees and managers.

As organizations look for new ways to obtain business value from their investments in enterprise systems, they are naturally seeking to integrate their systems externally with suppliers, customers, and other business partners by implementing what are often called interorganizational systems (IOS). Although many organizations have realized sizable benefits from IOS, it is clear that much more improvement can be achieved. The key to realizing additional benefits is to build IOS on the understanding that the benefits an organization receives from IOS depend, not only on its own implementation choices, but also on those of its business partners. This paper reviews the conditions promoting and inhibiting benefit realization from IOS, presents two examples of companies that successfully addressed those conditions, and examines why some organizations find it challenging to imitate these examples.

There is an increasing demand for efficient indexing techniques to support queries on large string databases. In this paper, a hybrid RAM/disk-based index structure, called the Hybrid Digital tree (HD-tree), is proposed. The HD-tree keeps internal nodes in the RAM to minimize the number of disk I/Os, while maintaining leaf nodes on the disk to maximize the capability of the tree for indexing large databases. Experimental results using real data have shown that the HD-tree outperformed the Prefix B-tree for prefix and substring searches. In particular, for distinctive random queries in the experiments, the average number of disk I/Os was reduced by a factor of two to three, while the running time was reduced in an order of magnitude.

Peer-to-peer (P2P) networks present the advantages of increased size of the overall database offered by a the network nodes, fault-tolerance support to peer failure, and workload distribution. Music file storage and exchange has long abandoned the traditional centralised server-client approach for the advantages of P2P networks. In this paper, we examine the problem of searching for similar acoustic data over unstructured decentralised P2P networks. As distance measure, we utilise the time warping. We propose a novel algorithm, which efficiently retrieves similar audio data. The proposed algorithm takes advantage of the absence of overhead in unstructured P2P networks and minimises the required traffic for all operations with the use of an intelligent sampling scheme. Detailed experimental results show the efficiency of the proposed algorithm compared to an existing baseline algorithm.

The World Wide Web is a popular broadcast medium that contains a huge amount of information. The web warehouse is an efficient and effective means to facilitate utilization of information on the Web. XML has become the new standard for semi-structured data exchange over the Web. In this paper, therefore, we study the XML web warehouse and propose an approach to the problems of change detection and warehouse maintenance in an XML web warehouse system. This paper has three major contributions. First, we propose an object-oriented data model for XML web pages in the web warehouse as well as system architecture for change detection and warehouse maintenance. Second, we propose a change detection method based on mobile agent technology to actively detect changes of data sources of the web warehouse. Third, we propose an incremental and deferred maintenance method to maintain XML web pages in the web warehouse. We compared our approach with a rewriting approach to storage and maintenance of the XML web warehouse by experiments. Performance evaluation shows that our approach is more efficient than the rewriting approach in terms of the response time and storage space of the web warehouse.

The set of groupware tools used during a distributed development process is usually chosen by taking into account predetermined business politics, managers’ personal preferences, or people in charge of the project. However, perhaps the chosen groupware tools are not the most appropriate for all the group members and it is possible that some of them would not be completely comfortable with them.

Continuous queries over data streams have gained popularity as the breadth of possible applications, ranging from network monitoring to online pattern discovery, have increased. Joining of streams is a fundamental issue that must be resolved to enable complex queries over multiple streams. However, as streams can represent potentially infinite data, it is infeasible to have full join evaluations as is the case with traditional databases. Joins in a stream environment are thus evaluated not over entire streams, but on specific windows defined on the streams. In this paper, we present windowed implementations of the traditional nested loops and hash join algorithms. In our work we analytically and experimentally evaluate the performance of these algorithms for different parameters. We find that, in general, a hash join provides better performance. We also investigate invalidation strategies to remove stale data from the window buffers, and propose an optimal strategy that balances processing time versus buffer size.

Data classification is an important problem in data mining. The traditional classification algorithms based on decision trees have been widely used due to their fast model construction and good model understandability. However, the existing decision tree algorithms need to recursively partition dataset into subsets according to some splitting criteria i.e. they still have to repeatedly compute the records belonging to a node (called F-sets) and then compute the splits for the node. For large data sets, this requires multiple passes of original dataset and therefore is often infeasible in many applications. In this paper we present a new approach to constructing decision trees using pre-computed data cube. We use statistics trees to compute the data cube and then build a decision tree on top of it. Mining on aggregated data stored in data cube will be much more efficient than directly mining on flat data files or relational databases. Since data cube server is usually a required component in an analytical system for answering OLAP queries, we essentially provide “free” classification by eliminating the dominant I/O overhead of scanning the massive original data set. Our new algorithm generates trees of the same prediction accuracy as existing decision tree algorithms such as SPRINT and RainForest but improves performance significantly. In this paper we also give a system architecture that integrates DBMS, OLAP, and data mining seamlessly.