ABSTRACT The asynchronous evolution of tests and code can compromise software quality and project longevity. To investigate the impact of test and production code co‐evolution, this study analyzes a large‐scale dataset of 526 GitHub repositories written in six programming languages: JavaScript, TypeScript, Java, Python, PHP, and C#. We focus on understanding how tests evolve throughout the software lifecycle and the frequency with which production and test code evolve in sync. By applying clustering algorithms and Pearson's correlation coefficient, we identify different patterns of test co‐evolution between projects. We found a significant correlation between high test co‐evolution and smaller development teams but no significant relationship with the frequency of different maintenance activities (corrective, adaptive, perfective, or multi). Despite this, we identified five distinct test evolution patterns, highlighting diverse approaches to integrating testing practices. This work provides valuable insights into the dynamics of test co‐evolution and its correlation in software maintainability.

Background: In software development, the identification of source code file experts is an important task. Identifying these experts helps to improve software maintenance and evolution activities, such as developing new features, code reviews, and bug fixes. Although some studies have proposed repository-mining techniques to automatically identify source code experts, there are still gaps in this area that can be explored. For example, investigating new variables related to source code knowledge and applying machine learning aiming to improve the performance of techniques to identify source code experts. Aim: The goal of this study is to investigate opportunities to improve the performance of existing techniques to recommend source code files experts. Method: We built an oracle by collecting data from the development history and surveying developers of 113 software projects. Then, we use this oracle to: (i) analyze the correlation between measures extracted from the development history and the developers’ source code knowledge and (ii) investigate the use of machine learning classifiers by evaluating their performance in identifying source code files experts. Results:First Authorship and Recency of Modification are the variables with the highest positive and negative correlations with source code knowledge, respectively. Machine learning classifiers outperformed the linear techniques (F-Measure = 71% to 73%) in the public dataset, but this advantage is not clear in the private dataset, with F-Measure ranging from 55% to 68% for the linear techniques and 58% to 67% for ML techniques. Conclusion: Overall, the linear techniques and the machine learning classifiers achieved similar performance, particularly if we analyze F-Measure. However, machine learning classifiers usually get higher precision while linear techniques obtained the highest recall values. Therefore, the choice of the best technique depends on the user’s tolerance to false positives and false negatives.

Having a full and healthy life is what most people want. This involves a set of domains such as physical and psychological health, social relationships, and environmental issues. Together, these domains can be used to characterize a person's Quality of Life (QoL). This desire for a healthy life becomes even more prominent as the world population ages. Thereby, new socioeconomic challenges arise, putting pressure on governments and industries to adopt technologies to improve QoL. One of the technologies that have been widely used is the Internet of Health Things (IoHT). However, developers still face challenges when proposing IoHT solutions, such as lack of interoperability, high volatility, difficulties in providing runtime adaptations, high development cost when it involves AI, and the absence of a semantic model for QoL data. Thus, this Ph.D. project aims to investigate the development of QoL-based IoHT systems to address these challenges. As a result, an IoT-Health Platform is expected to support this development and bring health professionals closer to this process.

In Brazil, a current health problem is the low capacity of meeting an increasing demand for medical services. As a result, some people have resorted to supplementary health care, which involves the operation of private health plans and health insurance. However, many health maintenance organizations (HMO) face financial difficulties due to unnecessary procedures, fraud or abuses in the use of health services. In order to avoid unnecessary expenses, the HMO began to use a mechanism called prior authorization, where a prior analysis of each user's need is made to authorize or deny the required requests. This work aims to study the influence of the use of textual features in automatic prior authorization evaluation, by using Text Mining, Natural Language Processing and Machine Learning techniques. Experiments were performed using several machine learning algorithms combined with textual features, increasing the performance of the automatic prior authorization. Results indicate not only the textual features influence to the evaluation of the automatic prior authorization process but also improved the prediction of the classifiers.

In a Health Plan, beneficiaries can cancel their contracts at any given time. For that reason, Health Insurance/Plan Providers (HIP) need to avoid optional contract cancellations to keep their financial operations stable. This work's main purpose is to develop an approach to predict the optional contract cancellation in a Private HIP and help them to prevent those cancelations.

The most common way of performing regression testing is by executing all test cases associated with a software system. However, this approach is not scalable since time and cost to execute the test cases increase together with the system’s size. A way to address this consists of prioritizing the existing test cases, aiming to maximize a test suite’s fault detection rate. To address the limitations of existing approaches, in this paper we propose a new approach to maximize the rate of fault detection of test suites. Our proposal has three steps: i) infer code components’ criticality values using a fuzzy inference system; ii) calculate test cases’ criticality; iii) prioritize the test cases using ant colony optimization. The test cases are prioritized considering criticality, execution time and history of faults, and the resulting test suites are evaluated according to their fault detection rate. The evaluation was performed in eight programs, and the results show that the fault detection rate of the solutions was higher than in the non-ordered test suites and ones obtained using a greedy approach, reaching the optimal value when possible to verify. A sanity check was performed, comparing the obtained results to the results of a random search. The approach performed better at significant levels of statistic and practical difference, evidencing its true applicability to the prioritization of test cases.

Software development, in many moments, is an exciting and challenging activity, but it can present itself as a boring endeavour in others. At the same time, the introduction of game elements into efforts such as the teaching of Software Engineering shows that real‐world activities can assemble game design elements and that it can make them more engaging. In this work, it is proposed the introduction of game design elements in software development, especially in the Scrum process. For this, elements are included to stimulate adherence to the prescriptions of the process, besides stimulating the execution of more activities by the team, positively impacting the productivity of a project. The authors present the idealised mechanics and the results obtained from the accomplishment of a case study in a software development team in a private company. Overall, the gamification applied to software development stimulated developers to perform their daily tasks, although this result did not generate strong evidence of increased productivity.

Fraud and abuse are two factors directly related to high health care costs, since they correspond to expenses that can be eliminated without prejudice to the quality of services provided. In Brazil, the health insurance companies implement a claim authorization process which assists in the detection of fraud and abuse. This process consists of a prior analysis of the services requested by providers, allowing them to detect patterns linked to fraud and abuse. This analysis is commonly performed manually, making the execution expensive and non-scalable. Health insurance companies have invested in the use of data mining and machine learning techniques to detect suspicious fraudulent patterns. However, the use of these techniques in claim authorization process is affected by the class imbalance problem, due to the fact that there are much more authorized service requests than unauthorized ones. This paper presents the investigation results of the effects of class imbalance in health insurance claims authorization domain. By means of an experiment, the performance loss of several classifiers was measured in different class distributions and also the performance recovery provided by treatment methods. The results show that the studied classification algorithms are affected differently by class imbalance. They also show that the recovery performance is lower the higher the class imbalance.

Software testing consists in the dynamic verification of the behavior of a program on a set of test cases. When a program is modified, it must be tested to verify if the changes did not imply undesirable effects on its functionality. The rerunning of all test cases can be impossible, due to cost, time and resource constraints. So, it is required the creation of a test cases subset before the test execution. This is a hard problem and the use of standard Software Engineering techniques could not be suitable. This work presents an approach for test case prioritization and selection, based in relevant inputs obtained from a software development environment. The approach uses Software Quality Function Deployment (SQFD) to deploy the features relevance among the system components, Mamdani fuzzy inference systems to infer the criticality of each class and Ant Colony Optimization to select test cases. An evaluation of the approach is presented, using data from simulations with different number of tests.

Software development is sometimes considered a boring task. To avoid this fact we propose an approach based on the incorporation of game mechanics into Scrum framework, in order to change its use to a more amusing task, by taking advantage of the gamification trend. Gamification is applied to non-game applications and processes, trying to encourage people to adopt them. This work shows a suggestion of Scrum gamification together with an evaluation of the proposed approach in a case study of a software house. The use of this concept can help the software industry to increase the team productivity in a natural way.

Computational Intelligence (CI) embraces techniques designed to address complex real-world problems in which traditional approaches are ineffective or infeasible. Some of these techniques are being used to solve several complex problems, such as the team allocation, building products portfolios in a software product line and test case selection/prioritization. However, despite the usefulness of these applications, the development of solutions based in CI techniques is not a trivial activity, since it involves the implementation/adaptation of algorithms to specific context and problems. This work presents Athena, a visual tool developed aiming at offering a simple approach to develop CI-based software systems. In order to do this, we proposed a drag-and-drop approach, which we called CI as a Service (CIaaS). Based on a preliminary study, we can state that Athena can help researchers to save time during the development of computational intelligence approaches.

Brazil has one of the largest private healthcare markets in the world. However, it appears that many indicators of our health insurance companies (including consultations, tests, hospitalizations, etc.) are well above the established international standards. This indicates that many appointments and tests are being carried out without need, generating unnecessary costs on businesses and making the service offered more expensive. In this paper, computational intelligence techniques were used to model the behavior of medical reviewers - professionals who assess whether medical requests should be allowed or not. For the generation of knowledge, a database from a nonprofitable health insurance company containing more than one million records collected since the year 2007 was used. Promising experimental results are presented, indicating that the techniques used can support medical reviewers.

Software Product Line (SPL) development is a new approach to software engineering that aims at the development of a whole range of products. One of the problems which hinders the adoption of that approach is related with the management of the products of the line. Additionally, the scope of a software product line is determined by the bounds of the capabilities provided by the collection of products in the product line. This introduces new challenges related to the scope problem. One of the main three different forms of scoping is the Product Portfolio Scoping (PPS). PPS aims at defining the products that should be developed as well as their key features. While this has an impact on the actual reuse opportunities, it is usually driven from marketing aspects. Defining a product portfolio by considering costumers satisfaction and cost aspects is a NP-hard problem. This work presents a hybrid approach, which combines fuzzy inference systems and the multi-objective metaheuristics NSGAII to support product management by generating portfolios of products, based in segments of users and the development cost of the assets of the SPL. Fuzzy inference systems are used to generate development cost of an asset by using coupling, number of code lines and cyclomatic complexity and also to estimate the quality of the products generated by the optimization module of our approach. The NSGA-II metaheuristic is used to search for products minimizing the cost and maximizing the relevance of the candidate products. The results show that the proposed approach is effective in proposing the best products in terms of relevance and cost of the assets.

The success of the team allocation in a agile software development project is essential. The agile team allocation is a NP-hard problem, since it comprises the allocation of self-organizing and cross-functional teams. Many researchers have driven efforts to apply Computational Intelligence techniques to solve this problem. This work presents a hybrid approach based on NSGA-II multi-objective metaheuristic and Mamdani Fuzzy Inference Systems to solve the agile team allocation problem, together with an initial evaluation of its use in a real environment.