Many educational problems are complex and there are no quick fixes with immediate answers to solve them. Evasion is one of these problems and deserves special attention, which justifies the need to carry out studies, research and critical reflections on the nuances that underlie, permeate and constitute approaches that can contribute to understanding and combating this phenomenon.The use of Information and Communication Technology brings people together, even if they are physically distant, but it can create a problem for those who do not have access to technological resources, which makes it an obstacle for people who do not have or have not developed certain skills.The technology resources that Distance Education uses, through Virtual Learning Environments, allow the construction of knowledge to occur in different spaces, with teachers and students developing activities in different places or times. Thus, the use of these educational technology resources available in Virtual Learning Environments will be addressed in this work from a mapping of educational resources used in undergraduate courses of the Open University of Brazil program at the University of Brasília. This research is characterized by being an exploratory case study and a qualitative approach was adopted, in which the data
analysis carried out identified the most used resources and related them to the dropout rate of the respective courses. The results obtained suggest that the diversity of resources can impact dropout rates.

Context: measuring user experience (UX) is essential to create value in digital transformation. Measurement allows you to identify future purchase intentions, user loyalty, and retention. The traditional way of measuring user experience using self-assessment has problems. Therefore, we need a more straightforward approach to allow us to measure user experience automatically.
Objective: implement and validate the UX evaluation model through a tool that automatically calculates the user experience rating of a digital product and presents positive, neutral, and negative points in using this product. The model included a work process, an exploratory experiment, and an application, which we call UXAPP.
Methods: we identified and selected the state-of-art related to emotion recognition with Artificial Intelligence in the context of user satisfaction. After that, we proposed and implemented the UX evaluation model. We developed the UXAPP application and conducted the exploratory experiment following the work process. We invite nine participants to perform four tasks each. We collect manual input data and capture emotions from the user's video and speech. Then, we analyze both data and generate a UX report. Finally, we compare the results obtained from UXAPP and the user input data.
Results: the UXAPP UX evaluation directly matches the user evaluation in 50\% of the tasks and gives a close answer in others 47.22\% of the results. Each UX element, like usability, affect, and user value, was evaluated and analyzed independently. The UXAPP usability analysis matches the user evaluation in 30.56\% of the tasks and gives a close answer in others 33.33\% of the results. The UXAPP affect analysis matches the user evaluation in 52.78\% of the tasks and gives a close answer in others 41.67\% of the results. UXAPP also identified 616 points of positive or negative sentiments in video and speech for all tasks. The UXAPP user value analysis matches the user evaluation in 36.11\% of the tasks and gives a close answer in 50.00\% of the results. We registered 5h 35m 31s of tasks’ duration. Based on our experience, we estimate that a person needed at least four times this task duration to process all the information presented in the UX report, including finding all sentiment points and distinguishing positive and negative sentiment peaks.
Conclusions: the possibilities of applying emotion recognition are countless in terms of contexts, techniques, forms, and components. Despite this, it was possible to develop UXAPP and validate the UX evaluation model to measure the user experience of a digital product through emotion recognition and Artificial Intelligence. We found that UXAPP analysis and user evaluation analysis are important and complementary. The main contribution of this work is the delivery of a validated model to automatically measure user experience and identify positive, neutral, and negative points in the use of a digital product. UXAPP can be used directly by the final user, making it possible to automate all the efforts to obtain a UX evaluation score and identify user interaction's positive and negative points. It can drastically reduce the costs involved in developing a user experience competence in an organization. This research also defines a measurement experimentation process that any organization or researcher can reproduce.

main goal:Mapping the relationships in which a higher education institution is inserted, from the perspective of systems dynamics;Complementarily, the following are identified as specific objectives of the research:Develop a bibliographic review related to the research topic, including specific literature, studies by specialized organizations, decrees, norms and legislation issued by educational regulatory bodies;Understand the risk factors linked to the momentary transition from face-to-face classes to remote ones;Understand the risk factors linked to the return to face-to-face classes;Identify the educational impacts during the transition, based on indicators of stricto sensu graduate students at the studied institution;

With the growing demand in the software development market and increasingly tight deadlines driven by agile methodologies, the associated risks to these software have increased in recent years. Cybercriminals exploit security, quality, and compliance vulnerabilities to commit cybercrimes against companies, resulting in financial losses and damage to the organizations’ reputation. Therefore, it is essential for companies involved in the development and supply of such software to understand how to identify and prioritize issues that require immediate attention.
In the literature, it was observed that existing studies address software risks in isolation and do not provide a consolidated view of these risks, lacking a model to assist in decisionmaking to prioritize which risk and which part of the software need to be addressed urgently.
In response to this challenge, this study aimed to understand the risks involved and explore the methods, techniques, and tools available for validating these risks in market software. After identifying the risks, methods, techniques, and tools were applied to the software, validating the presence of these risks. Upon confirming the existence of the risks, the FAHP multicriteria decision support method was used to assist in risk classification, determining which part of the software and which risk should be prioritized first.
The results indicated that, among the nine software modules, the Web module with 34.57%, combined with the Vulnerability risk with 50.35%, needs to be prioritized. This decision support model emerges as a contribution to decision-making, especially in the field of software engineering.

This dissertation investigates the effectiveness of non-linear machine learning (ML) models in predicting the Brazilian consumer price index (IPCA). By leveraging advanced ML techniques such as Random Forest, AdaBoost, Extreme Gradient Boosting (XGBoost), and Gradient Boosting, we aim to enhance the accuracy of inflation forecasting within an emerging market context. Spanning from August 2010 to January 2024, our analysis utilizes a comprehensive dataset comprising 156 predictors. To optimize model performance, we employ recursive feature elimination (RFE) with ElasticNet, isolating the 30 most influential predictors.

Among the models examined, Gradient Boosting emerges as the most effective, exhibiting superior accuracy indicators. Notably, we enhance the interpretability of this winning model through the Shapley Value, an Explainable Artificial Intelligence (XAI) technique. By elucidating the individual contributions of variables, Shapley Value mitigates the "black box" nature of ML predictions, offering valuable insights into the dynamic interactions between predictor variables essential for strategic decision-making within institutions.

This dissertation underscores the potential of integrating sophisticated ML methodologies with traditional macroeconomic tools to refine forecasting capabilities. It also points towards promising avenues for future research, such as exploring deep learning and hybrid models. Moreover, our findings extend beyond academia, offering practical insights for navigating complex economic landscapes.

The Secretariat of Justice and Citizenship of the Federal District has partnerships with the third sector, commonly called Civil Society Organizations, which must issue an accountability report, at the end of the projects, for analysis by the public entity. The present work aims to use Risk Management to mitigate possible failures in the analysis of the accountability processes of Civil Social Organizations, in order to resolve the following issue: Public Administration can mitigate risks related to divergent measurements of the agreed results in terms of partnership, between public administration and Civil Social Organizations - CSOs, so that the perception of the weights of the criteria considered mainly by decision-makers, are similar between approval, approval with reservation or rejection of accounts? For this purpose, the external and internal context, criteria, identification, analysis and assessment of Risks were established, as well as the treatment action defined within the scope of the research. For this purpose, tools such as Failure Modes and Effects Analysis, Current Reality Tree and Hierarchical Analytical Process for Group Decision were used with calculations using the Bayesian theory before applying the Individual Aggregation of Judgments. The results showed that the use of the Bayesian theory rationalized the methodology for applying the Analytic Hierarchy Process tool, together with the Individual Aggregation of Judgments and that among the weights of the criteria in group consensus, the Social Impact proved to be irrelevant even being the main purpose of the work of Civil Society Organizations, which is revealed in another risk. When analyzing the weight of the criteria in the alternatives, it was observed that the criterion Cost is the one that most interferes and Efficiency least interferes in the specialist’s evaluation, as for approval, what interferes most is Effectiveness and what interferes least is Cost . 

The Central Bank of Brazil (BCB) is responsible for assisting customers and users of the National Financial System (SFN) with complaints against products and services offered by its supervised entities and ensuring that all demands are appropriately addressed. Each concern is manually handled and classified as “proceeding” or “unfounded” based on a preliminary analysis of the facts described by the customer, the alleged entity’s reply, the attached documents, and according to its compliance with current regulations. However, dealing with an increasing demand with the available resources has turned out to be an unprecedented challenge, being currently impossible to handle all issues accordingly. In this context, the BCB has developed an automated solution to filter complaints more likely to be classified as “proceeding”, thus driving the human activities of examination, analysis and judgment. The present study aims to propose a new classifier for this task with better performance than the current model. To this end, deep learning approaches were explored, differing from the traditional method previously employed. As a result, an experimental classifier was tested, based on a hierarchical structure, and combining Bidirectional Encoder Representations from Transformers (BERT) with Bidirectional Long Short-Term Memory (BiLSTM) to generate unique contempt representations of both the citizen’s complaint and the entity’s reply. In cross-validation, this solution reached an average performance of 71.41%, based on the area under the precision-recall curve (PRAUC), exceeding the current BCB model by 0.96 percentage points, and performing 71.92% on the test dataset. If the proposed classifier was adopted to drive the task of handling complaints, we estimate that approximately 90.23% of the proceeding demands could be identified by evaluating only 60% of the total amount. In addition, multimodal strategies were tested to combine the described textual representations with tabular features of the original classifier. However, when compared to the previous proposed solution, the achieved gain with the multimodal strategies did not reach a statistically significant outcome.

The gamification of non-playful activities gained ground in the industrial, service, and educational environment, and the literature points out that by introducing game elements into an environment, activities can be more attractive, fun and more enjoyable. Combining game techniques to motivate and engage people has proven to be a powerful tool for managers. Military Police ́s from Federal District is the police corporation responsible for the ostensible policing and maintenance of public order in the federal capital. The Intelligence Activity is used to advisory on strategic, tactical, and operational decisions. This work presents the Gamified Intelligence Reporting Process (PGRI) to be utilized in Military Police Intelligence System. PGRI was constructed using Octalysis gamification framework base. Built using Intrinsic Motivation Inventory questionnaires and document analysis submited to statistical concepts. The PGRI pursue to improve the quality of intelligence reports with the application of gamification and is fed back by an iterative cycle of planning through the process evaluation results. It presents a literature review about the concepts of games and gamification, Intelligence Activities and Public Security and a summary about the Military Police of the Federal District.

This work aims to propose a computational tool for Risk Management related to Human Fatigue, based on the Fatigue Risk Management System - Sistema de Gerenciamento do Risco à Fadiga (FRMS) approach, applied to Primeiro Centro Integrado de Defesa Aérea e Controle de Tráfego Aéreo (CINDACTA I), in order to minimize the risks related to fatigue. To this end, exploratory, descriptive, applied research was carried out, with a qualitative and quantitative approach. Documents from the Air Force Command, national and international legislation, application of a questionnaire, brainstorming with specialists in the areas of investigation and prevention of aeronautical accidents, and air traffic control were used as a data collection technique. As a result of this study, the perception of operators who work in operational bodies was collected. Based on the literature, the professionals answered a questionnaire with 44 objective questions directly related to the Human Fatigue (FH) theme, which can influence Operational Performance (OD). To tabulate the resulting data, partial least squares structural equation modeling (PLS-SEM) was applied, which revealed that according to the response of 149 operators, the conceptual model elaborated explains the operational performance in 34.7%, the dimension reaction time in 31.2% and state of alert in 29.9%. In the validation stage, the computational solution was presented to specialists and analysts, through meetings, aiming at validating the Sistema de Gerenciamento de Riscos da Fadiga (SGRF) and evaluating the feasibility of using the tool by the institution. The study revealed that, in order to improve the operational performance of air traffic controllers, actions should be taken regarding sleep quality and reaction time. Finally, a prototype of a computational tool developed, tested, and validated by the Organization’s specialists, called the Sistema de Gerenciamento de Riscos da Fadiga (SGRF), was presented, aiming to contemplate all the essential stages of the fatigue risk management process for the Air Traffic Control Officer (ATCO), according to the Fatigue Risk Management System (FRMS) approach.

For some time now, (inter)national researchers have been studying dropout rates in higher education courses, classifying them into two types: students who drop out of the university, and students who drop out of higher education. Both situations cause damage to the institution, students, and society at large. Starting in 1995, with the creation of ANDIFES, studies began to be more frequent in Brazil. This commission developed reports that analyzed graduation rates, retention, and dropout rates in undergraduate courses at Brazilian universities. Institutional dropout was one of the study subjects, described as a student leaving their original course without completing it. The University of Brasília (UnB), considering the issues surrounding student dropout, has created mechanisms to increase student retention in undergraduate courses. The objective of this work was to develop and evaluate an analytical model that allows the use of academic data for predicting dropout rates in face-to-face undergraduate courses. A Systematic Literature Review was conducted to identify the factors that impact dropout rates and define indicators that can be extracted from UnB's academic systems. It also helped in selecting algorithms/tools to support the analysis. The main result of the systematic review was the identification of 29 factors used by researchers, where average score, gender, and course grades were the most commonly used ones. Regarding tools, Regression, Decision Tree, and Neural Network were the most frequently used algorithms. Based on this preliminary result, the Undergraduate Analysis Model (MAGRA) was created, which utilizes existing indicators in UnB's academic systems in conjunction with machine learning tools to predict students at risk of dropout. The research was conducted in an environment that encompasses two scenarios. The first was developed at the Faculty of Gama (FGA), which served as a prototype for MAGRA's creation, and the second at UnB. In the testing stages, where only the courses taken by students were considered, it was shown that the number of times a student takes a course can be an indicator of their difficulty in completing it within the designated time frame. By employing the new variables, there was an increase in the number of valid models and, consequently, an increase in the analyzed courses and classes, resulting in a higher number of predictions. This situation was observed during the studies conducted at the Faculty of Gama and the University of Brasília. To improve early identification of students with dropout characteristics, it is necessary to create feedback mechanisms from course coordinators, introduce new systems, improve data quality, and adjust algorithm parameters.

The Brazilian Federal Court of Accounts organizes its work by processes and, throughout their life cycle, each of them usually contains from tens to hundreds of legal documents. Each document easily reaches a few dozen pages. The number of processes and documents only tends to grow over time, which generates a huge amount of material for reading and with a very rich content, but difficult to consume, as it takes considerable time to read each process. The processes are usually read to verify if they have relevant content for any fiscalization or procedural instruction in progress. In addition to the high cost of reading a process, part of this content is discarded by the auditor because it is not linked to their current work, which generates a waste of time in this activity. To improve the efficiency of this process, we proposed in this work the development of an automatic text summarization solution using machine learning applied to natural language processing. This solution will use the abstractive summarization approach applied to long documents and with legal content, using models that are state-of-the-art in the task and based on transformers with linear attention mechanism. The solution will be made available as an Web Apllication with a microservice for better integration with applications that make up the auditor’s work process. The summaries generated by the models will be evaluated mainly by metrics that focus more on the semantics of the generated text and, as a result, will have a better adherence to the desired content. The user will provide feedback on the generated summaries and they will be used to feed back the model later.

Advances in communication and computing technologies have boosted the use of the internet, affecting the life of society and playing an important role in business. The banking industry has followed this evolution, offering essential services through the internet to millions of customers. In addition, competition between banks that offer services through the internet has grown more and more. In this scenario, it is observed that the User Experience (UX) is an important factor in the perception of quality by the customers of these institutions, even influencing the choice between one bank and another. As the software architecture is an important element that affects usability, it is essential to ensure that usability and user experience criteria are supported when planning this architecture. Therefore, this work proposes an architecture for banking systems based on concepts of reactive systems, providing for integration with services and heterogeneous data sources to improve the UX, with an interface that adapts to the user based on their behavior and characteristics. For this reason, adaptive interfaces tend to offer a better user expe- rience, as they seek to adapt to the needs and desires of users. There are several ways to render adaptive applications, this work follows the micro-frontends model, as it gener- ates smaller, more cohesive and sustainable code bases, making it possible to work with autonomous teams and different technologies. These characteristics are important, espe- cially for larger corporations, which support larger groups of professionals and teams. The architecture proposed by this work uses the Reinforcement Learning algorithm, added to the Monte Carlo Tree Search (MCTS) algorithm and the Deep Learning technique, to create adaptive applications. Finally, a robust backend system has to be developed to extract, store and process user data, so the microservices’ architecture was adopted, which allows applications to be formed by small, cohesive and independent services.

The use of the information system brings constant benefits to those who use it, due to the increase in data generated. Grouping this information can be useful for daily activities in any work environment, and even in personal life. When talking about information system, we do not necessarily talk about computerized system, but organization, either through some automated platform or simply in a manual file. Along with the use of SI, there are risks that involve its development and that may affect its perceived usefulness. Thus, the objective of this dissertation was to propose steps to minimize the risk factors that impact the use of use of information systems delivered by the Undersecretariat of Information Technology of the Ministry of Agriculture and Livestock. Initially, it was important to search the literature for the risk variables that the authors understand to be most relevant. Subsequently, the structural equations model was applied to relate these variables and thus identify the factors with the greatest impact on the use of the IS. Regarding the practical implications, a questionnaire was first applied, where 71 responses were collected. The main results are: Perceived Utility influences the use of IS in 38.24% of cases; and Ease of Use influences Perceived Utility 15.21% of the time. In addition, the importance-performance map (IPMA) was generated to enrich the results of the PLS-SEM model and assist in the prioritization of the variables of greatest relevance and impact. In order of high importance and high performance, the indicators perceived utility and risks will be prioritized. Moreover, the moderation of the model revealed the variable Age as a point of attention because it is a variable of impact on the perceived usefulness of information systems.

Investments in financial operations can generate impacts on the social, economic, and environmental sectors that were not incorporated into the final price of their contracts. Such impacts are called externalities and it is understood that they can produce positive or adverse effects. The analysis of the credit risk associated with the socio-environmental concerns the possibility of economic activities generating socio-environmental impacts or damages capable of generating financial losses that compromise the ability to pay the financing, as well as the co-responsibility of the financial agent for the recovery of socio-environmental damage. In the case of climate risk, the golf venture is directly or indirectly exposed to extreme weather events capable of jeopardizing its cash flow and ability to pay to finance. Thus, there was a need for measurement instruments capable of evaluating the behavior of externalities in relation to the credit risk of operations. The objective of this study is to assess the impact of incorporating climate risk measurement into financial operations in the strategic process of providing credit for operations. For this, a measurement of the climate risk was carried out and a methodology was applied —— to assess the importance of the variables that compose it and infer their ability to influence the credit analysis. As a result of the present study, measurement of climate risk of operations and assessment of climate risk in relation to variables were provided.

In an increasingly dynamic environment, the speed of decision-making has become a differential, especially in a scenario that has many obstacles, such as the government. Interests from different spheres (municipal, state and federal) and different actors (political parties, benches, mobilization of popular groups), must be consolidated in the most assertive decision for society. Thus, methods that guarantee the decrease of subjectivity and increase the consolidation of different points of view are necessary. Thus, the objective of this study is to demonstrate a decision-making model through the use of the hierarchical analytical process (AHP) in partnership with the Promethee II method, applying optimization through Risk Simulator to prioritize the information technology demands of the Ministry Federal Public. This is an exploratory research with a qualitative-quantitative approach using multicriteria analysis and optimization. As a preliminary result of the study, the criteria Support, Strategic and Interested Parties were established, with the respective approximate weights of 65%, 28% and 7%. Eight alternatives were defined, where the most prominent was "Technical knowledge of the ICT team" with an approximate weight of 24.5%. The combination of criteria and alternatives applied during the research resulted in the demand for Security being the highest priority and the demand for BI being the lowest priority among the 12 selected. It is also up to the computational application of the Risk Simulator tool to carry out the optimized prioritization according to restrictions defined by senior management.

Digital transformation encompasses the use of technologies and their impact on organization’s business models, resulting in changes in products, organizational structures or process automation. The adoption of innovative resources can contribute to the improvement of work processes, especially regarding the reduction of repetitive tasks, errors and recurring problems. In this context, digitization is related to business development, with processes being supported by digital solutions or functions; while process automation is applied with the use of digital technologies to increase efficiency, reducing steps in workflows and enabling the automation of repetitive human tasks. The people management area of the Federal Institute of Education, Science and Technology of Brasília (IFB) faces the challenges of modernizing the traditional methods used in its processes management. The area uses electronic means to carry out administrative processes, in addition to structuring technological platforms. The plurality of solutions with low or no integration causes repetitive activities of low complexity, with excessive consumption of time. The objective of this work is to define a digital transformation approach for repetitive processes in the people management area of the IFB. The adopted research methodology is descriptive with the use of the Case Study technique. The approach will be based on Robotic Process Automation, an emerging technology for process automation, which was selected from the execution of a systematic literature review to identify techniques, tools, solutions and implementations used in digitalization with emphasis on process automation

Jurisprudence refers to the set of repeated decisions on a given subject, constituting a type of judicial precedent. Within the scope of the Federal Audit Court (TCU), the body responsible for exercising external control of the Federal Public Administration, jurisprudence represents the consolidated interpretations of the rules applicable to the financial and operational supervision of the public accounts of the Union’s bodies and entities. Since the elaboration of jurisprudence is defined based on a grouping of similar rulings, it is important to develop automated tools that assist the specialists responsible for this activity. However, this is a challenging task for the area of computing, due to the specificities of the vocabulary present in the texts of the rulings and the massive volume of data to be processed. Therefore, it is necessary to develop scalable, effective and efficient approaches that have low computational cost. This work presents the study and implementation of some approaches for representing these textual documents, both at the word level and at the concept level. As a contribution, a new approach called BoC-Th (Bag of Concepts with Thesaurus) was proposed, which generates weighted histograms of concepts defined based on the distance of the words in the document to their respective similar term within a thesaurus. This approach allows us to emphasize words with greater meaning in the context, thus generating more discriminative vectors. Experimental evaluations were carried out comparing the proposed approach with traditional approaches for document representation. The proposed method obtained superior results among the techniques evaluated for recovering jurisprudential documents. BoC-Th increased average accuracy compared to traditional approaches, including the original BoC (Bag of Concepts), while also being faster than traditional BoW, BM25, and TF-IDF representations. The proposed approach contributed to enriching an area with peculiar characteristics, providing a resource for retrieving textual information more accurately and quickly than other techniques based on natural language processing.

The Centro de Coordenação de Operações Móvel (CCOp Mv) is a strategic project of the Brazilian Army that is under development, with the aim to support a Large Operational Command in war and non-war situations. CCOp Mv will be composed of different technologies, systems and resources, in order to achieve its objective. Among its main functionalities can be highlighted the acquisition and provision of information of field events, which in addition to directing the construction of situational awareness, represent the essential parameters for the decision-making process of authorities, in military operations. Despite the technological advances experienced today, the nature of current conflicts allows for the perception of the relevance of obtaining data through human observers. However, simply obtaining data is not considered sufficient for the formation of specific knowledge about a given domain, requiring the construction of a structured information base, which allows the recovery of knowledge in an optimized way. In this context, the proposal of this research is to provide an Information Architecture aimed at obtaining data through field observers, in addition to providing generated knowledge, having as parameters the organization and informational reports structuring, employing ontologies for its semantic representation. Furthermore, the construction of an ontology prototype is being proposed, in order to support the developed information architecture, since the domain of military events has a considerable specificity. As contributions can be highlighted the provision of an Information Architecture, which aims to assist in the optimization of obtaining information in military operations, as well as the prototype of ontology that was developed, which could be used as a basis for the development of other applications within the Brazilian Army.

The Intelligence Department of the Military Police of São Paulo State (CIPM/PMESP) has been evolving their legacy systems, which are currently in a microservices architecture. This architecture splits an application into small services, which are implemented independently, with their own deployment unit. The microservice architecture can bring benefits, however, it also presents challenges, especially regarding security aspects. Therefore, it brings a need to explore knowledge about security issues in microservices, especially in authentication and authorization aspects. In addition, the characteristic of the intelligence activity must necessarily reach a series of principles, such as compartmentalization, which aims to restrict access to certain information only to professionals who need to know. This reinforces the need to verify whether the current environment is adequate in terms of authentication and authorization, to ensure compliance with this principle. This work aims to propose and evaluate the security of a microservices-oriented solution to perform authentication and authorization in the Intelligence Systems of the Military Police of the State of São Paulo. To this end, a Systematic Literature Review is carried out to check what are the main challenges in authentication and authorization in the microservice architecture, the mechanisms that deal with such challenges and the open-source technologies that implement such mechanisms. With the SLR findings, a validation survey is carried out, verifying whether such findings are observed in the industry, to confirm the practical use of what was found, in addition to seeking to verify new answers that were not found in the SLR. Finally, an implementation is carried out in the PMESP intelligence environment with the open-source technology that implements the mechanisms found, applying security tests in the current and new environment, to verify if the applied mechanisms improved the security of the application.

otnet is a network of infected computers, which are remotely controlled by a cyber-
criminal, called botmaster, whose objective is to carry out massive cyber attacks, such
as DDoS, SPAM and information theft. Traditional botnet detection methods, usually
signature-based, are unable to detect unknown botnets. Behavior-based analytics has held
promise for detecting current botnet trends, which are constantly evolving. Considering
that Botnet attacks on the IT infrastructure of the Brazilian Army’s Mobile Operations
Coordination Center (CCOp Mv) may harm the success of operations, through theft of
sensitive information or even causing interruption to critical CCOp Mv systems, this dis-
sertation proposes a botnet detection mechanism based on network flow behavior analisys.
The main objective is to propose an additional layer of cyber protection to the CCOp
Mv IT infrastructure. The technique used to detect botnets was recently developed and
it is called Energy-based Flow Classifier (EFC). This technique uses inverse statistics to
detect anomalies and has an important characteristic which is its easy adaptation to new
domains. Due to this characteristic, EFC is a promising technique for detecting unknown
botnets. EFC uses only benign data to infer the detection model and classifies as mali-
cious any flow that deviates from the normal traffic pattern learned during model training.
Thus, in addition to flows related to botnet activities, other types of malicious activities
may be detected, making further verification necessary to identify the type of malicious
activity detected. Two heterogeneous datasets, CTU-13 and ISOT HTTP were used to
evaluate the efficiency of the model and the results were compared with several traditional
algorithms. Preliminary results show that EFC presented good results when tested in the
same domain and the tests performed in different domains show that the EFC manages
to maintain stable results, regardless of the domain, unlike the other models tested.

In disaster response operations, a great deal of coordination of various agents and institutions is required. The use of tactical and operational software is of great importance for the success of the operation. However, operations can take place in adverse locations with scarce Internet access resources. This limitation can generate problems in the provision of services, since such applications use remote processing, which is typically performed a distant datacenter. In this context, this research project proposes an architecture that explores the concepts of fog computing called MFog, with a focus on application resilience, this architecture can be used in places where, despite the existence of structures with computational resources close to the end user, access to the Internet is limited (similar to what happens in these operations). For this, minimal architecture components and orchestration mechanisms are defined based on container application migration models. The results obtained from a case study indicate that the proposed architecture can be resilient to the problems faced in the scenarios of these operations.