Olavo Holanda

Steps, techniques, and technologies for the development of intelligent applications based on Semantic Web Services: A case study in e-learning systems

Semantic Web Services domain has gained special attention in academia and industry. It has been adopted as a promise to enable automation of all aspects of Web Services provision and uses, such as service creation, selection, discovery, composition, and invocation. However, the development of intelligent systems based on Semantic Web Services (SWS) is still a complex and time-consuming task, mainly with respect to the choice and integration of technologies. In this paper, we discuss some empirical issues associated with the development process for such systems and propose a systematic way for building intelligent applications based on SWS by providing the development process with steps, techniques and technologies. In addition, one experiment concerning the implementation of a real e-learning system using the proposed approach is described. The evaluation results from this experiment showed that our approach has been effective and relevant in terms of improvements in the development process of intelligent applications based on SWS.

A framework for building web mining applications in the world of blogs: A case study in product sentiment analysis

Recently there has been much interest in electronic commerce applications that use data mining techniques to explore datasets in the social media context. However, most of the applications have already been developed in an ad hoc manner, mainly, due to the lack of adequate tools, yielding difficulties in customizing applications and requesting high time consuming for constructing and maintaining these applications. This work addresses these problems and proposes a software framework for building Web mining applications in the blog world. The architecture of the proposed framework combines the use of blog crawling and data mining algorithms, in order to provide a complete and flexible solution for building general-purpose Web mining applications. The framework flexibility allows some important customizations, such as the construction of adapters for reading text from different blogs, and the use of different pre-processing techniques and data mining algorithms. In order to improve the efficacy of information extraction from blogs, ontology is used in the blogs description. For this, there are software agents responsible for tracking and indexing blogs related to a specific tag and for mining blog datasets. Moreover, web services are used for encapsulating existing tools and maximize reuse. This framework has been instantiated in order to be applied for helping the blog users to effectively find out relevant information in the blog world. The focus of this paper is on describing the novel software architecture of the general framework (blog crawling and data mining) providing detailed information about the data mining sub-framework, which uses the semantic web services technology for automating service composition and consists on the main research contribution. A case study of an e-commerce application for analyzing the users sentiment regarding specific products is reported and its results considers the effort reduction when creating a web mining application by using the proposed integrated frameworks and existing data mining tools, as well as a qualitative analysis related to quality aspects of the developed application, such as the evolution impact.

JOINT: Java ontology integrated toolkit

In the past few years, the use of ontologies for creating more intelligent and effective application has increased considerably. This growth is due to the fact that ontologies attempt to provide semantics to the data consumed by machines so that they can reason about this data. However, developing complex ontology-based applications is still difficult and time-consuming because the existing tools do not provide a simple and unified environment for developers. Most of these tools only provide data manipulation using RDF triples, complicating the development of applications that need to work with the object orientation paradigm. Furthermore, tools that provide instances manipulation via object orientation do not support features such as manipulating ontologies, reasoning over rules or querying data with SPARQL. In this context, this work proposes a framework and a tool for supporting the efficient development of ontology-based applications through the integration of existing technologies. Furthermore, we also define a methodology to use this tool efficiently. In order to evaluate the benefits of our work, a controlled experiment with eight developers (unfamiliar with ontologies) was performed to compare the proposed tool, JOINT, with another one, Jastor/Jena, frequently used by the community. The results suggest that our tool helps novice developers to create ontology-based applications faster and with few errors in the code. In addition, a real educational application with 10 ontologies, more than 200 ontology concepts (classes) and more than a million triples is already using the proposed tool successfully.

Educational resources recommendation system based on agents and semantic web for helping students in a virtual learning environment

This work proposes an agent-based system for helping students in the knowledge building process by providing them with recommendations of papers and blogs. The interaction starting point involves a blog context focused on a specific domain. Then, the teacher posts a topic in this blog and students express their opinions about this topic. Afterwards, the system monitors the interaction process aiming to mine the discussion to extract important attributes and information. Then, the recommender system retrieves relevant material to support dynamic interactions in the learning environment. Thus, two situations were defined one, being that the system helps the students by finding and recommending relevant and related blogs. The other, being that the system helps students by recommending papers from a digital library to provide them with formal opinions about the target topic, aiming at enriching their learning experiences. The proposed system with a case study and empirical results are described.

Ontology-based software product line for building semantic web applications

The Software Product Lines (SPL) has proved very effective in building large-scale software. However, few works seek to adjust the approach of software product line to applications in the context of semantic web. This is because applications in this context assume the use of semantic services and intelligent agents. As a result, it is necessary that there are assets that provide adequate interoperability both semantic services and intelligent agents. In this sense, it is proposed in this paper the use of ontologies for the specification of entire a project of a SPL. With this, it can be a sufficiently formal specification that can be interpreted by both software engineers and computational algorithms.

An Ontology-Based Model for Driving the Building of Software Product Lines in an ITS Context

The Intelligent Tutoring Systems (ITS) have been used daily to support education in various domains. For this reason, fast and easy construction of ITSs are fundamental requirements. In this sense, a Software Product Line has been used accordingly for building Intelligent Tutoring System families. However, the construction of such family systems is still a hard and complex task which involves the representation and manipulation of different knowledge sources with distinct artifacts. To alleviate these issues, this paper proposes an ontology-based model for driving the building of software product lines in an ITS Context. It also provides a case study describing the construction of an ITS in the programming domain. In addition, an evaluation is presented aiming to show the feasibility of the proposed model. The main conclusion is that this model reduces the effort and the complexity in the construction of such systems.

A semi-automatic system to evaluate the performance and scalability of ontology persistent APIs

Abstract The international Semantic Web community has produced several tools, known as API – application programming interfaces, to help development teams to create applications using semantic technologies. One important family of tools developed to date is related to the manipulation of instances in ontology-based applications. There are two main approaches for manipulating ontology instances in currently developed APIs: (a) using RDF (resource description framework) triples or (b) using classes in object-oriented programming (OOP). In APIs based on RDF triples the development teams need to understand how ontology works in the RDF layers in order to manipulate data for each triple in the application code. Such an approach is highly desired to develop clean applications based strongly on Semantic Web Technologies. Nevertheless, it requires development teams to change their programming paradigm and master those technologies. In APIs based on OOP developers can manipulate data at object level in order to make application development simple and flexible using the conventional programming paradigm. Because of that, several developers have shifted from RDF triples to OOP. Although several tools have been developed to manipulate ontologies at object level, many have not been adequately evaluated for performance and scalability. In effect, there is an urgent demand for effective research results regarding metrics and systems to evaluate the performance and scalability of ontology persistent APIs. As a result, these systems should facilitate the evaluation of persistent APIs to developers. This study thus aims to provide a system for developers to evaluate ontology persistent APIs at object level. In order to validate the feasibility of our system, we have conducted an experiment by considering two solutions used by community of ontology persistence.

An object triple mapping system supporting detached objects

The development and use of ontologies in creating more intelligent and effective applications have increased in recent years. This rise in popularity is due to the potential of ontologies in providing semantics for the data consumed by machines, allowing them to reason on these data. Providing sophisticated tools that improve the development of applications based on RDF and OWL can further accelerate the adoption of the Semantic Web. In this context, some types of ontology tools have been proposed, for instance, ontology editors, ontology reasoners, triple store tools, RDF frameworks and object triple mapping (OTM) systems. In particular, OTM systems (e.g., AliBaba, KOMMA, JOINT and so on) are of utmost importance for developing ontology-based applications through the object-oriented paradigm. They map RDF triples into data objects, enabling developers to handle RDF triples as objects in object-oriented applications. In a similar way, developers of relational database applications have been using object-relational mapping (ORM) systems, for example Hibernate, to map relational databases into objects as well as to manage the object states. The main object state that Hibernate provides is the detached one; however, except for Empire, current OTM systems do not provide detached objects. The great difference between these two types of objects is that the former one has its life cycle independent of the underlying database connection, while the latter one is bound to the connection. In this way, this paper presents an object triple mapping system that supports detached objects, named JOINT-Detached (JOINT-DE). By using JOINT-DE, developers of ontology-based applications can (i) use the objects coming from the triple store as objects of the business model, (ii) use such objects as data transfer objects (DTOs) between subsystems, and (iii) use of less resources (triple store connections) to load object properties. To evaluate the proposed tool, we first describe how we conducted an experiment to empirically compare our tool with another OTM widely used by the community. We measure performance and memory usage regarding create, retrieve and update operations. Our results indicate that: (1) JOINT-DE has better performance results in all three operations; (2) JOINT-DE has an advantage over AliBaba with respect to the memory usage in the retrieve operation; (3) AliBaba uses less memory than JOINT-DE in the update operation; and (4) we could not find statistically significant differences between the tools regarding memory usage in the create operation. Finally, to illustrate the benefits of the proposed system, a case study of a real application is presented, outlining the architectural limitations of the application using an existing OTM in the literature, as well as showing positive results to the use of JOINT-DE.

Towards an Agent-Based Approach for Automatic Generation of Researcher Profiles Using Multiple Data Sources

In this paper, we study the problem concerning the automatic generation of user profiles as a function of multiple data sources. Hence, we introduce an agent-based approach for collecting and integrating data from multiple sources, such as Linked In, Google Scholar and Lattes, in order to automatically generate researcher profiles. We performed a simple evaluation involving seven users who have data in all the used sources. Preliminary, the results show the feasibility of our approach.