Flavia Cristina Bernardini

Artificial Neural Networks Ensemble Used for Pipeline Leak Detection Systems

The physical and operational properties of pipelines vary greatly. There is thus no universally applicable method, external or internal, which possesses all the features and the functionality required for a perfect leak detection performance. The authors of this paper know quite well that traditional methods, in a low uncertainty environment, overcome artificial intelligence methods of leak detection systems. If one considers the real world as a creator of uncertainties, neural networks and fuzzy systems emerge as important promising technologies for the development of leak detection systems. In this work, we propose a method for constructing ensembles of ANNs for pipeline leak detection. The results obtained in our experiments were satisfactory.Copyright

Transparency in practice: using visualization to enhance the interpretability of open data

Urban data is gradually being opened to the public. Tools for exploitation, analysis and discovery of new knowledge in large data sets are the key to enable citizens to make sense of such large amount of data. The purpose of this work is to analyze how data analysis associated with visualization techniques in different levels can lead to the improvement of the interpretability of open data. With the support of machine learning techniques, these visualizations may improve pattern identification in urban data sets. To guide our discussion, a case study was conducted analyzing socioeconomic data released by the Chicago city government. We discussed the use of different visualizations in this scenario, tailored for univariate, bivariate and multivariate analysis. We also performed an evaluation of the different forms of visualization proposed in this work. We could observe that allowing the user to explore open urban data using some specific visualizations may lead to more effective data interpretation.

A parallel fipa architecture based on GPU for games and real time simulations

The dynamic nature and common use of agents and agent paradigm motives the investigation on standardization of multi-agent systems (MAS). The main property of a MAS is to allow the sub-problems related to a constraint satisfaction issues to be subcontracted to different problem solving agents with their own interests and goals, being FIPA one of the most commonly collection of standards used nowadays. When dealing with a huge set of agents for real time applications, such as games and virtual reality solutions, it is hard to compute a massive crowd of agents due the computational restrictions in CPU. With the advent of parallel GPU architectures and the possibility to run general algorithms inside it, it became possible to model such massive applications. In this work we propose a novel standardization of agent applications based on FIPA using GPU architectures, making possible the modelling of more complex crowd behaviours. The obtained results in our simulations were very promising and show that GPUs may be a choice for massively agents applications. We also present restrictions and cases where GPU based agents may not be a good choice.

Building a Reference Model and an Evaluation Method for cities of the Brazilian Network of Smart and Human Cities

Smart Cities theme has evolved in the last years, leading the cities to implement initiatives related to technical aspects to improve quality of life. Another focus on this theme is how to measure the added value of these initiatives to the population. The Brazilian Network of Smart and Human Cities (RBCIH) was created in order to join both approaches in Brazil, putting together members from academy, private initiative and local (municipality) government. Nowadays, RBCIH is composed by 350 Brazilian cities (in a universe of 5570 cities), indicating that a long-term work has to be executed by RBCIH in Brazil. Two purposes of RBCIH are presenting good practices and evaluating how smart and human a city is. In this work, a methodology is presented for constructing a Brazilian reference model and an evaluation method for smart cities, adherent to the Brazilian reality.

City Ranking Based on Financial Flux Indicator Clustering

Nowadays, many cities around the world adopted different definitions of Smart Cities. Many organizations proposed collections of indicators for evaluating how smart the cities are. One problem related to applying these indicators is how to compare fairly the cities, considering that they are quite different in some countries or regions, like in Brazil. Therefore, grouping similar cities should be interesting, though there are different methods for grouping them. This work proposes a method for grouping cities based on their financial features, supported by clustering and regression techniques. This work also present an evaluation of the proposed method using real data from cities located in Rio de Janeiro state in Brazil. The results show the feasibility of the method, although obtaining the specifically type of used data is yet a challenge.1

Brazilian Social Mood: The Political Dimension of Emotion

Brazil faces a major economic and political crisis. Millions of people joined anti-government protests across the country. Social media sites are a way for some people to vent their emotions without feeling self-conscious. Thus, emotion mining on social media can be viewed as effective tool to conduct Presidential approval rating. This research aims to automatically recognize emotion in texts extracted from social media in Brazilian Portuguese (PT-BR). The ultimate goal is knowing how emotions influence a writer of a text in choosing certain words and/or other linguistic elements. In this research, we perform keyword-based approaches using affect lexicon and a Support Vector Machine and Naive Bayes algorithms.

An Expert System Based on Parametric Net to Support Motor Pump Multi-Failure Diagnostic

Early failure detection in motor pumps is an important issue in prediction maintenance. An efficient condition-monitoring scheme is capable of providing warning and predicting the faults at early stages. Usually, this task is executed by humans. The logical progression of the condition-monitoring technologies is the automation of the diagnostic process. To automate the diagnostic process, intelligent diagnostic systems are used. Many researchers have explored artificial intelligence techniques to diagnose failures in general. However, all papers found in literature are related to a specific problem that can appear in many different machines. In real applications, when the expert analyzes a machine, not only one problem appears, but more than one problem may appear together. So, it is necessary to propose new methods to assist diagnosis looking for a set of occurring fails. For some failures, there are not sufficient instances that can ensure good classifiers induced by available machine learning algorithms. In this work, we propose a method to assist fault diagnoses in motor pumps, based on vibration signal analysis, using expert systems. To attend the problems related to motor pump analyses, we propose a parametric net model for multi-label problems. We also show a case study in this work, showing the applicability of our proposed method.

Todas as palavras da sentença como métrica para um sumarizador automático

The purpose of this work is to present an automatic summarizer that uses as a metric the number of words into a sentence to define the text authors pragmatic profile. Using the number of words as a metric, the original text is classified according to its temporal measures and textual composition, which is based on its formality. Also, these features are parameters to the summary generation that indicate the compression level. This work uses traditional methodologies of automatic summarization and compares these results to results obtained with our proposal.

Using Geocoding and Topic Extraction to Make Sense of Comments on Social Network Pages of Local Government Agencies.

Social networks have become an important channel for exchanging information and communication among citizens. Text mining, crowdsourcing and data visualization are some approaches that allow the information and knowledge extraction from texts in comment formats, exchanged between citizens in social networks. This movement can be indirectly used as a bias for popular participation, gaining prominence in the construction of smart cities. The objective of this work is to present a method that geocodes citizens’ comments made on posts in Social Network Pages of Local Government Agencies, and extracts the most frequent topics present in these comments. In order to validate our method, we implemented a web system that implements the steps of the proposed method, and conducted a case study. The tool, and consequently the steps of the presented method, was evaluated by four software developers, which indicated that the tool was easy to use, new knowledge could be extracted from it, and some interesting improvements were pointed out by them.

An instrument for evaluating open data portals: a case study in Brazilian cities

Open data portals are fundamental tools for governments to achieve public transparency. Hence, several countries around the world have issued Freedom of Information (FOI) laws and acts, imposing that city administrations develop their own open data portals to publish local open data. However, many cities, specially in developing countries, lack the financial resources even to invest in basic IT services. These cities quite often do not have qualified people in open data best practices, transparency and IT solutions for constructing these portals, specially the middle and small-sized ones. There are many works pointing out guidelines and general problems in open data portals. However, there is a lack of an unified instrument for evaluating and suggesting good practices in constructing open data portals considering both open data portal requirements and FOI access laws and acts. In this work, we propose an instrument for evaluating open data portals that gathers functionalities present in platforms that support open data portals construction and recommendations from FOI access law. Specifically, we used recommendations from Brazilian Information Access Law, as it is an instance of FOI access recommendation. We evaluated and validated our instrument assessing portals of 26 Brazilian cities. Our case study shows that the instrument is effective for showing many problems in these portals, and also reveals the current scenario in these cities.