Identifying and reducing customer churn have become a priority for financial institutions seeking to retain clients. Our research focuses on customer churn rate analysis using advanced machine learning (ML) techniques, leveraging a synthetic dataset sourced from the Kaggle platform. The dataset undergoes a preprocessing phase to select variables directly impacting customer churn behavior. SMOTETomek, a hybrid technique that combines oversampling of the minority class (churn) with SMOTE and the removal of noisy or borderline instances through Tomek links, is applied to balance the dataset and improve class separability. Two cutting-edge ML models are applied—random forest (RF) and the Light Gradient-Boosting Machine (LGBM) Classifier. To evaluate the effectiveness of these models, several key performance metrics are utilized, including precision, sensitivity, F1 score, accuracy, and Brier score, which helps assess the calibration of the predicted probabilities. A particular contribution of our research is on calibrating classification probabilities, as many ML models tend to produce uncalibrated probabilities due to the complexity of their internal mechanisms. Probability calibration techniques are employed to adjust the predicted probabilities, enhancing their reliability and interpretability. Furthermore, the Shapley Additive Explanations (SHAP) method, an explainable artificial intelligence (XAI) technique, is further implemented to increase the transparency and credibility of the model’s decision-making process. SHAP provides insights into the importance of individual features in predicting churn, providing knowledge to banking institutions for the development of personalized customer retention strategies.

Climate change is a phenomenon that is sometimes denied or trivialized. However, in recent years, we have faced extreme phenomena such as fires, floods, excessive temperatures, etc. which affect our physical and mental condition and the environment, often leading to significant material damage. To understand these problems and highlight the meteorological and phenomenological changes encountered in the last decade, time series were web-scraped and analyzed from several open data sources: weather news broadcast in Romania, air quality, temperature, etc. The extraction and organization of data recorded between 2009 and 2023 are formulated as a framework that can be reproduced and replicated to continue the monitoring. The exploratory analysis of the categorical and numerical data highlights intricate patterns and correlations within meteorological conditions across regions and seasons. From temperature trends to air quality fluctuations, the study underscores the dynamic interplay of weather phenomena, paving the way for informed forecasting and deeper climate research. At the same time, data processing includes Latent Dirichlet Allocation, K-prototype clustering analysis, in addition to K-means clustering with dimensional reduction techniques, all of which are employed to further reveal the extreme phenomena in news and regions with higher occurrence. Therefore, in this paper, we propose a data processing framework for multiple datasets and analytics, extracting valuable information on climate change and identifying the exposed regions to extreme phenomena.

The extreme phenomena have been increased recently in frequency and intensity causing numerous damage that cannot be neglected by residents, local authorities and social media. More European countries are experiencing a growing number of occurrences, such as floods, powerful winds, storms, icing, and unusual temperature fluctuations. Particularly, the year 2023 has recorded the highest temperatures in the history of humanity. In this research, we compile a dataset that combines news reports with numerical data pertaining to weather conditions and air quality at the historical region level in Romania. We examine the news and recorded data spanning the years from 2009 to 2023 using anomaly detection and clustering techniques to compare the results. Specifically, we employ Isolation Forest and Autoencoders to identify anomalies within the data that are further clustered to analyse the detection process. We explore the occurrence frequency and duration of daily simultaneous extreme weather events over the years, conducting statistical tests like the Mann–Kendall test to discern trends in the extreme phenomena. The findings reveal statistically significant increasing trends in the incidence of heatwaves, storms and floods. When we set the Mean Squared Error (MSE) threshold to 95%, both methods detect nearly 16% of the anomalies, and this figure rises to over 25% when the MSE threshold is set to 90%. An analysis of anomalies at the regional level indicates that most anomalies are detected in the Transylvania and Muntenia regions, while the Banat region experiences the lowest level of anomalies.

Cardiovascular diseases (CVDs) are the leading cause of death globally. Detecting this kind of disease represents the principal concern of many scientists, and techniques belonging to various fields have been developed to attain accurate predictions. The aim of the paper is to investigate the potential of the classical, evolutionary, and deep learning-based methods to diagnose CVDs and to introduce a couple of complex hybrid techniques that combine hyper-parameter optimization algorithms with two of the most successful classification procedures: support vector machines (SVMs) and Long Short-Term Memory (LSTM) neural networks. The resulting algorithms were tested on two public datasets: the data recorded by the Cleveland Clinic Foundation for Heart Disease together with its extension Statlog, two of the most significant medical databases used in automated prediction. A long series of simulations were performed to assess the accuracy of the analyzed methods. In our experiments, we used F1 score and MSE (mean squared error) to compare the performance of the algorithms. The experimentally established results together with theoretical consideration prove that the proposed methods outperform both the standard ones and the considered statistical methods. We have developed improvements to the best-performing algorithms that further increase the quality of their results, being a useful tool for assisting the professionals in diagnosing CVDs in early stages.

Climate change is an aspect in our lives that presents urgent challenges requiring innovative approaches and collaborative efforts across diverse fields. Our research investigates the growth and thematic structure of the intersection between climate change research and machine learning (ML). Employing a mixed-methods approach, we analyzed 7521 open-access publications from the Web of Science Core Collection (2004–2024), leveraging both R and Python for data processing and advanced statistical analysis. The results reveal a striking 37.39% annual growth in publications, indicating the rapidly expanding and increasingly significant role of ML in climate research. This growth is accompanied by increased international collaborations, highlighting a global effort to address this urgent challenge. Our approach integrates bibliometrics, text mining (including word clouds, knowledge graphs with Node2Vec and K-Means, factorial analysis, thematic map, and topic modeling via Latent Dirichlet Allocation (LDA)), and visualization techniques to uncover key trends and themes. Thematic analysis using LDA revealed seven key topic areas, reflecting the multidisciplinary nature of this research field: hydrology, agriculture, biodiversity, forestry, oceanography, forecasts, and models. These findings contribute to an in-depth understanding of this rapidly evolving area and inform future research directions and resource allocation strategies by identifying both established and emerging research themes along with areas requiring further investigation.