Breast cancer is one of the most prevalent and life-threatening diseases among women worldwide, making early and accurate detection crucial for effective treatment. This study presents a novel and optimized breast cancer classification system using machine learning models enhanced through advanced hyperparameter tuning techniques and statistical validation methods. The publicly available Breast Cancer Wisconsin (Diagnostic) dataset was utilized and is undergoing preprocessing steps, including PCA-LASSO for feature selection and scaling for dimensionality reduction. The classifiers employed include Random Forest, Support Vector Machine (SVM), Gradient Boosting, and Logistic Regression, which were further refined using GridSearchCV, RandomizedSearchCV, and Optuna, with 3-fold cross-validation implemented to ensure robust evaluation of model performance. The novelty of this work lies in integrating PCA-LASSO for feature selection with Optuna-driven optimization and ensemble learning strategies to achieve superior accuracy. A comprehensive evaluation of the proposed models was performed using key classification metrics such as precision, recall, F1-score, and accuracy, with the Optuna-optimized ensemble model achieving the highest accuracy of 99.423%. Statistical validation of the results was conducted using ANOVA tests, confirming the significance of the improvements achieved through hyperparameter optimization. Additionally, feature importance scores for Random Forest and Gradient Boosting provide insights into the most influential factors in the classification process. This work demonstrates the efficacy of combining PCA-LASSO feature selection, ensemble learning, and Optuna optimization, validated through 3-fold cross-validation, offering a novel and reliable approach to interpretable breast cancer detection systems.

Transmitted through mosquito bites, malaria is a disease that poses a serious threat to life and is caused by parasites. Its impact on the health of infants, children, and pregnant women globally is substantial, often leading to child malnutrition and contributing to a rise in infant mortality rates. According to the 2020 World malaria report published by WHO, there was significant increase in 14 million more cases and 69000 more deaths as compared to 2019. The emergence of insecticide and drug resistance in mosquitoes has led to setbacks in malaria control effects. Hence, early diagnosis and timely treatment are essential for effective management and control of the disease. However, traditional diagnostic methods are time-consuming, costly, and require trained personnel. Machine learning techniques can help to overcome these challenges by predicting and classifying malaria risk using clinical information. In this research paper, we propose a malaria risk prediction model using machine learning techniques based on clinical information. Our study utilizes a data set of patient records from a malaria endemic area, including demographic information, clinical symptoms, laboratory test results, and treatment history. We used machine learning classification models like logistic regression, decision tree classifiers, gaussianNB, random forests, and extratree classifiers, to build and evaluate prediction models. We used feature selection approaches to determine the most useful characteristics for malaria risk prediction. The proposed machine learning-based approach has several potential applications in clinical practice. For example, the approach can be used to identify high-risk individuals and allocate resources for early detection and treatment. The approach can also support public health officials in developing targeted intervention strategies to reduce malaria transmission. In conclusion, this study demonstrates the effectiveness of machine learning techniques in malaria risk prediction using clinical information. The proposed approach can potentially improve early detection and treatment of malaria, ultimately reducing the burden of the disease.

Federated learning (FL) and artificial intelligence (AI) are the two most cutting-edge and well-liked technologies in the field of intelligent healthcare. The medical industry has depended on centralized agents sharing data that has not been processed. Therefore, this method still has a lot of problems and flaws. FL is the process of training statistical models on mobile devices from a central location using decentralized data while keeping the privacy of the devices safe. Privacy-preserving FL methods have a lot to offer to learning-based healthcare systems, especially in the light of patient privacy requirements. This manuscript used the PRISMA guideline to conduct a systematic review. The goal of the paper is to give a thorough understanding of the architecture and difficulties associated with FL’s growing use in the healthcare sector. Publications related to FL in healthcare were gathered, categorized, and examined using a systematic review methodology. A total of 186 papers were selected from IEEE, Elsevier, Arxiv, ACM, MDPI, and WOS databases. 38 duplicates were removed during the identification phase. Furthermore, 49 publications were excluded after assessing their titles and keywords. Twenty-three of the 99 papers that satisfied the full-text eligibility criteria were excluded due to insufficient relevance and inadequate focus on quantitative evaluations. Consequently, 76 papers were chosen for full-text review. The terms "FL", "Healthcare", and "aggregation" were used in the English-language publication of the inclusion measures. The paper aims to give a thorough analysis of the architecture, aggregation techniques, and difficulties associated with the growing usage of FL in the healthcare industry. We discuss the aggregation techniques, implementation frameworks, and design to analyze and assess FL’s capabilities along with robust aggregation approaches. However, the aggregation technique selected determines the final global model’s quality and dependability; therefore, choosing the correct method is essential. A thorough grasp of the various model aggregation strategies and their advantages and disadvantages is necessary, as FL becomes more prevalent across multiple disciplines. Additionally, we thoroughly examined FL’s applicability in the healthcare sector, because in the FL model, training occurs at the client site only. FL excels in protecting privacy because it does not require centralized data aggregation, which presents serious privacy problems. Researchers may use this survey to improve and create novel aggregation methods, while practitioners can use it to choose the best approach for their FL applications.