Course: Master´s in Informatics
Institution/Location: Federal University of the State of Rio de Janeiro (UNIRIO) - Rio de Janeiro - Brazil
ABSTRACT: The studies related to water resources have great relevance in many areas such as irrigation, water supply and power generation. The efficient use of these resources depends on many factors, like the correct estimation of certain variables related to the hydrological cycle, such as evapotranspiration. However, the most precise models currently applied for estimating evapotranspiration require variables that are not always available or are too complex to obtain in some regions, due to the lack of measuring instruments. In these cases, the precision of the evapotranspiration estimative is decreased, compromising its validity depending on the context. This research consisted in the application of Data Science techniques over meteorological data provided by the Brazilian National Institute of Meteorology (INMET), in order to generate a model for estimating evapotranspiration, using a "data-driven" approach. As a Data Science project, this research had high level of interaction with a domain expert from Hydrology area. This interactive process was necessary for definition of the research question, experimental scenarios and for results evaluation, generated by the successive runs of the Data Science lifecycle used in this research. Through interaction with the domain expert, the main objective of this research was defined to simplify the current methods for evapotranspiration estimation, without loss of precision in relation to the historical results. In order to automate the experimental runs, we developed a software program that supports all the steps of the Data Science lifecycle to enable the reproducibility of the experimental results. After successive runs of the experiment with scenarios defined together with the domain expert, we found a model that fits the goals defined in the first step of the lifecycle. Finally, for results analysis by the expert domain, graphs were generated to compare the results of different scenarios, as well as maps with layers of the Brazilian biomes and climate types, aiming to identify possible patterns among results and vegetation and climate type.
Course: Specialization in EAD Planning, Implementation and Management
Institution/Location: Fluminense Federal University (UFF) - Niterói - Brazil
Institution/Location:: XXX Brazilian Symposium on Databases (SBBD) - National Laboratory of Scientific Computing (LNCC) - Petrópolis - Brazil
ABSTRACT: The evaporation is defined as the sum of the evaporated water with the water transpired from the vegetation. It has great importance in irrigation, that represents 75% of global water consumption and, therefore, its estimate is crucial to better management of this resource. Currently, this estimate is made by mathematical methods, using data obtained from measuring stations, whose absence prevents the use of these methods. This initial study aims, from the data available at measuring stations, apply the KDD process to identify whether there are patterns in the estimation in places with similar characteristics, which could be used for similar locations without measuring stations.
Institution/Location: XI Brazilian Symposium on Information Systems (SBSI) - Goiânia - Brazil
ABSTRACT: With the growing volume of data in various areas such as Hydrology, there is a need for using information systems to aid in handling such data. This article is a report of an experiment that used knowledge discovery techniques to estimate an important component of the hydrological cycle: evapotranspiration. The experiment reported in this article was done with weather data and showed that some algorithms, such as M5P, present good results when compared to historical data of the estimated evapotranspiration.
Institution/Location: PICMET 2010 Conference - Phuket - Thailand
ABSTRACT: In times of big business competition, technology can be a very important tool for companies to get an advantage over their competitors. One example is a Radio-Frequency Identification (RFID) technology is getting more investments from companies around world. In Brazil, this technology is getting more investments from companies, universities and the government. This paper will present the real situation about RFID technology in Brazil and some case studies of RFID deployment in Brazilian companies from some sectors, such as aircraft industry, military and auto based industry. The main objective of this paper is to map RFID technology in Brazil, showing our advances and our difficulties. As a result, research on RFID deployments can be used in other research projects. Another result is that companies can analyze these case studies to get a better idea about their own RFID deployment projects.
Institution/Location: University of Campinas (UNICAMP) - São Paulo - Brazil
ABSTRACT: The use of software agents is increasingly common. Characteristics of some agents, such as mobility, are Fundamental for application in some areas. In technology RFID, the use of software agents can This paper aims to present a review of the work done On agents applied in RFID technology.