Paulo Marcelo Tasinaffo

Using GQM Hypothesis Restriction to Infer Bayesian Network Testing

By definition, the scope of a Bayesian Network uses a complementary technique to restrict the modeling reach. In this paper, the used restriction technique was the Goals, Questions, and Metrics (GQM). The hypothesis to be tested relates cause and effect conditional probabilities in a software test phase of a manufacturing production line. The Bayesian Network concept is related to the specific concept of a Directed Non Cyclic Graph (DNCG), where each one of its nodes represents a random discrete variable and is illustrated by directed arcs of cause and effect relationships between variables. A Bayesian Network is a graphical artifact which restricts problems, incorporating data structures. The major contributions of this paper are conceptualization and implementation of a methodology for using a GQM hypothesis restriction to infer Bayesian network testing with the Netica Bayesian Networks ® computer software.

A Methodology for Assessment Database According to ISO 9126

A methodology for implement, design and test, and an instrument for quality inspection complying with the ISO 9126 standard was developed by the Research Group on Software Engineering, GPES, and TATA Consultancy Services at the Aeronautics Institute of Technology, ITA. GPES is a research group based on ITA. The methodology and the instrument use the Causal Analysis and Resolution, CAR, process area, based on the approach Goal, Question, indicator, and Measures (GQ(i)M). The main contributions were: a) the creation and implementation of indicators, funded on metrics, for a decision process, which are based on the construction of a quality assessment methodology for a software product, in a model of a information system; and b) a prototype of an instrument for quality, which involves the approach GQ(i)M in the definition of quantitative and qualitative metrics, by means of a Microsoft Excel spreadsheet. The methodology represents the sophistication of the Causal Analysis and Resolution technique for gain identification, using database techniques. In this article the methodology was applied to two different case studies.

Implementing fuzzy logic to simulate a process of inference on sensory stimuli of deaf people in an e-learning environment

A novel approach to education aimed at deaf students, based on computing that performs individualized instruction on the domain of programming languages is presented. This approach is fully implemented and evaluated in an educational application model, called model of Mental Architecture Digitized—AMD. In particular, performs user modeling by dynamically identifying and updating a students knowledge level of all the concepts of the domain knowledge. The concept of AMD is based on fuzzy cognitive maps (FCMs) that are used to represent the dependences among the domain concepts. AMD uses fuzzy sets to represent a students knowledge level as a subset of the domain knowledge. Thus, it combines fuzzy theory with the overlay model. Moreover, it employs a novel inference mechanism that dynamically updates user stereotypes using fuzzy sets. It should be noted that the overlay model and stereotypes constitute two widely used methods for user modeling. The gain from this novel combination is significant as a student level of knowledge is represented in a more realistic way by automatically modeling the learning or forgetting process of a student with respect to the FCMs and thus, the system can provide individualized adaptive advice. The transmission and retention of knowledge rests on the cognitive faculty of the concepts linked to it. The repeatability of your applications builds a solid foundation for Education, according to behavioral standards set. This cognitive ability to infer on what we observe and perceive, regarded as intrinsic human beings, does not depend of their physical capacity.

An Agile Developed Interdisciplinary Approach for Safety-Critical Embedded System

Accidents and crises, whether climatic, economic, or social are undesirably frequent in everyday lives. In such situations, lives are sometimes lost because of inadequate management, lack of qualified and accurate information, besides other factors that prevent full situational awareness. The goal of this work is to report on an academic conceptualization, design, build, test, and demonstration of computer systems, to manage critical information, during hypothetical crises. During the development of an academic system in the second Semester of 2015 at the Brazilian Aeronautics Institute of Technology, the following challenges occurred: strict specifications, agile methods, embedded systems, software testing, and product assessment. Also, some quality, reliability, safety, and testability measurements have been used. At that time, an Interdisciplinary Problem-Based Learning (IPBL) was performed, adding hardware technologies of environment sensors, Radio Frequency Identification (RFID), and Unmanned Aerial Vehicles (UAVs). Software technologies were used for cloud-based web-responsive platform and a mobile application to geographically manage resources at real-time. Finally, the ANSYS® SCADE (Safety-Critical Application Development Environment) was employed to support the embedded and safety-critical portion of this system.

A Formal Control Model for Risks Management within Software Projects

This paper presents a proposed Formal Control Model(FCM) using a Colored Petri Net (CPN) and an inspection form for risks management within a software project. The basis for this model was the risk areas of the Capability Maturity Model Integration for Software Development (CMMI-DEV). The integration of risk elements from a formally defined quality model using a graphical and mathematical modeling tool has provided risks management. On the context of a Management Information System (MIS), a FCM prototype was developed to reduce human inference dependences, supporting organizational goals to track critical points for decision makers. The major contribution of this paper was the FCM conceptualization and application. The proposed model was applied to a project within the financial department of an enterprise CMMI level 5. It was able to identify, control, and manage risks of software development using a SG concept of CMMI risk applied to certain other CMMI PAs. At the end, a successful case study was performed involving the two experiments of Project Planning(PP) and Risk Management (RSKM). Their assessments have shown that after the proposed FCM execution, PENDING activities were completely fixed.

Process Model for Decision Making Using GQM approach

In anticipation of providing a software product quality, organizations select points to be measured in developing this product but do not have a question if they really are those points that should be measured. This work aims to propose a set of indicators that show how a deterministic error affects the business of the organization. They use the approach Goal-Question-Metric (GQM) which allows the analysis of the goals of the organization questions, and from these questions develop metrics that consider the variables that impact the business of organization. Measurements are performed when justified by a clearly defined goal and the indicators are generated if they are strongly related to the business objectives of this organization. After these indicators, is a report indicating what steps the organization should take to ensure that errors do not occur further. This report is a Technical Report, which enables decision-making more assertions by the organization. After the development of these indicators, they form the basis of historical errors and historical bases of metrics.

Systematic and Formal Approach to get a Domain Specific Language

This paper presents a systematic approach applied over State Machine (since it is a wide know model and easy to be used to formal specification) to improve the domain analysis procedure, besides been out of scope of this paper, this approach can also helps to improve the enterprises business process as well. The motivation which leads to the this paper is how to got a Domain Specific Language (DSL) that is completely correspondent to a Domain Analysis sharing both the same business rules. This is a very important property to be achieved since a DSL must be used to help the codding procedure in a specific domain, so, it must be a direct relation over them and this relation is explored in this paper. It has a briefly discussion over the need for formalization procedures concluding that too much formalization can be a problem and lack of it can also be, so, formal transformations can be performed at mark point (as baselines or any other mark that can be defined) bringing important contributions to the rigor of the model improving it.

Health Care Information Systems: A Crisis Approach

During the 1st Semester of 2017, at the BrazilianAeronautics Institute of Technology (Instituto Tecnologico de Aeronautica, ITA), a successful Interdisciplinary Problem-Based Learning (IPBL) experience took place. At that time, almost 30 undergraduate and graduate students from three different courses within just 17 academic weeks had the opportunity of conceptualizing, modeling, and developing a Computer System based on Big Data, Internet of Things, and other emerging technologies for governmental organizations and private sectors. The purpose of this system was to aggregate data and integrate actors, such as Patients, Hospitals, Physicians, and Suppliers for decision making processes related to crises management involving events of health systems, such as epidemics, that needs to manage data and information. Differently from other existing products from Universities, Research Centers, Governmental Agencies, Public and/or Private companies, this product was developed and tested in just 17 academic weeks, applying the Scrum agile method and its best practices available in the market. This experience was stored in a Google site and implemented as a Proof of Concept (PoC). It represents just one example of how to address the old problems of teaching, learning, and developing complex intelligent academic computer projects to solve health system problems, by collaboratively using the Scrum agile method with Python or Java, Spark, NoSQL databases, Kafka, and other technologies. The major contribution of this paper is the use of agile testing to verify and validate an academic health system case study.

Using Correct-by-Construction Software Agile Development

Disasters and crises, whether climatic, economic, or social are undesirably frequent in everyday lives. In such situations, lives are lost mainly because of inadequate management, lack of qualified and accurate information, besides other factors that prevent full situational awareness, including software failures. The goal of this paper is to report the agile conceptualization, design, build, and demonstration of a computerized system, containing correct-by-construction software, to safely manage critical information, during alerts or crises situations. On this research, the following challenges and requirements were tackled: formal specifications, aerospatial-level reliability, agile development, embedded systems, controlled testability, and product assessment. An Interdisciplinary Problem-Based Learning (IPBL), involving a Scrum of Scrums Agile Framework was adapted for managing the cohesive, productive, and collaborative development team of around 100 undergrad and graduate students remotely working. In addition, the following hardware technologies, for supporting the software development were used: environment sensors, Radio Frequency Identification (RFID), and Unmanned Aerial Vehicles (UAVs). Other software technologies were also used, as well cloud-based web-responsive platforms and mobile applications to geographically manage resources at real-time. Finally, the ANSYS® SCADE (Safety-Critical Application Development Environment) was employed to support the embedded and correct-by-construction module of this system, according to Model-Driven Architecture (MDA) and Model-Driven Development (MDD).

Using Monte Carlo method to estimate the behavior of neural training between balanced and unbalanced data in classification of patterns

This paper proposes to develop a model-based Monte Carlo method for computationally determining the best mean squared error of training for an artificial neural network with feedforward architecture. It is applied for a particular non-linear classification problem of input/output patterns in a computational environment with abundant data. The Monte Carlo method allows computationally checking that balanced data are much better than non-balanced ones for an artificial neural network to learn by means of supervised learning. The major contribution of this investigation is that, the proposed model can be tested by analogy, considering also the fraud detection problem in credit cards, where the amount of training patterns used are high.