Technical Debt (TD) is one of the main obstacles to the quality and sustainability of software in the long term, accumulating "interest" that increases maintenance costs and reduces the productivity of development teams. Among the strategies to manage TD, refactoring stands out for its ability to improve the internal structure of the code without changing its external behavior, contributing to improving the readability, maintainability, and extensibility of the software. This study analyzes the impact of refactoring on TD using metrics extracted from SonarQube and RefactoringMiner at the class level. TD is represented by the "sqale_debt_ratio" and stratified in quartiles to model its levels. The analysis focuses both on the presence or absence of refactoring and on the influence of specific types and levels of applications. Evaluating the role of refactoring is fundamental to understanding how these practices can reduce TD, improve software quality, and optimize maintenance. To ensure the robustness of the models, metrics directly related to TD were removed and collinearity-based selection techniques were applied. The validation of the approach was performed on six open-source software systems available on GitHub. The predictive models, built with machine learning techniques, were analyzed with SHAP (Shapley Additive exPlanations) to identify the most relevant features. The results of the study are promising and demonstrate how the adoption of targeted refactoring practices can effectively reduce TD. The combined approach of prediction and explainability helps to fill a gap in the literature and offers practical guidance for managing TD and optimizing refactoring practices in software systems.

Artificial intelligence and Machine Learning represent a fundamental role in the medical field today, even in the case of neurode-generative diseases such as Parkinson’s disease; particularly in non-invasive diagnostics. This work presents a machine learning-based approach for the early detection of Parkinson’s disease through voice recordings. Additionally, explainability techniques are applied to highlight key vocal features associated with the condition, supporting clinical interpretation and future integration in digital healthcare systems. The study aims to contribute to the timely diagnosis of Parkinson’s, providing an in-depth understanding of the disease to open up new opportunities for a more personalised and precise approach to therapies, representing a first step toward the future of predictive medicine.

Parkinson’s Disease (PD) is a neurodegenerative condition primarily affecting the elderly but also occurring in younger individuals. It is caused by a progressive loss of nerve cells in the brain’s substantia nigra that release dopamine, essential for controlling movements. Dopamine deficiency results in symptoms affecting both motor and non-motor functions, which vary among individuals. Diagnosis relies on clinical symptoms and medical history, often supported by brain scans, as there is no specific diagnostic test available. Diagnosis is challenging due to vague initial symptoms resembling other conditions. Current research indicates that AI can significantly enhance data and image analysis, aiding in the diagnosis and monitoring of PD progression. To this aim, this study proposes a hybrid model allowing the integrated use of clinical data and single photon emission computed tomography images of a patient to predict the presence of the disease. The approach consists of a combination of two types of neural networks, an LSTM for clinical data and a CNN for images. The validation is performed on a widely validated dataset belonging to the Parkinson’s Progression Markers Initiative, from which the data recording visits of 1,814 patients were extracted. The obtained results are interesting and useful to address further investigations.

Today, several chest diseases are on the rise and these are often diagnosed through the use of chest X-rays, a common and economical clinical test to perform. This work uses a machine learning approach for the detection of thoracic diseases using chest X-ray reports and involves leveraging algorithms and models to analyze medical imaging data for the presence of various conditions affecting the chest area. Our main goal is to create a predictive model based on textual reports released by radiologists, with the use of Natural Language Processing. The proposed approach aims to facilitate the examination of textual reports written by radiologists, to predict the onset of diseases in patients. Specifically, reports generated by radiologists are meticulously processed and reviewed using the GloVe and LSI models. This analysis allows you to identify the presence of diseases and provides insights into the specific thoracic pathology. The results obtained through the implementation of our approach (accuracy above 96% for the best model) underline the good performance and potential of the developed predictive model.

Parkinson's is a neurodegenerative disease, with a slow but progressive evolution, which involves some main functions such as the control of movements and balance. Symptoms vary by patient and include motor factors such as tremors and stiffness as well as non-motor symptoms such as cognitive impairment. Its diagnosis is not easy, so it is becoming increasingly necessary to assist doctors in identifying and predicting the disease. Artificial intelligence takes up this challenge and this work proposes a new approach to predict the onset of the disease and monitor patients. The experimentation involved the use of different classification algorithms. The proposed methodology was validated on a large ad hoc data set by compiling data collected by the Parkinson's Progression Markers Initiative (PPMI). Specifically, the study compares the results of the classification taking into consideration only the characteristics belonging to the motor sphere, or those of the non-motor sphere, with the aim of understanding which characteristics are more significant for the identification of the disease. In this regard, a multi-stage feature selection was conducted and SHAP was used to make the model explainable.