Cardiovascular disease (CVD) makes our heart and blood vessels dysfunctional and often leads to death or physical paralysis. Therefore, early and automatic detection of CVD can save many human lives. Multiple investigations have been carried out to achieve this objective, but there is still room for improvement in performance and reliability. This study is yet another step in this direction. In this study, two reliable machine learning techniques, multi-layer perceptron (MLP), and K-nearest neighbour (K-NN) have been employed for CVD detection using publicly available University of California Irvine repository data. The performances of the models are optimally increased by removing outliers and attributes having null values. Experimental-based results demonstrate that a higher accuracy in detection of 82.47% and an area-under-the-curve value of 86.41% are obtained using the MLP model, unlike the K-NN model. Therefore, the proposed MLP model was recommended for automatic CVD detection. The proposed methodology can also be employed in detecting other diseases. In addition, the performance of the proposed model can be assessed via other standard data sets.

No abstract

The detection and diagnosis of Breast cancer at an early stage is a challenging task. With the increase in emerging technologies such as data mining tools, along with machine learning algorithms, new prospects in the medical field for automatic diagnosis have been developed, with which the prediction of a disease at an early stage is possible. Early detection of the disease may increase the survival rate of patients. The main purpose of the study was to predict breast cancer disease as benign or malignant by using supervised machine learning algorithms such as the K-nearest neighbor (K-NN), multilayer perceptron (MLP), and random forest (RF) and to compare their performance in terms of the accuracy, precision, F1 score, support, and AUC. The experimental results demonstrated that the MLP achieved a high prediction accuracy of 99.4%, followed by random forest (96.4%) and K-NN (76.3%). The diagnosis rates of the MLP, random forest and K-NN were 99.9%, 99.6%, and 73%, respectively. The study provides a clear idea of the accomplishments of classification algorithms in terms of their prediction ability, which can aid healthcare professionals in diagnosing chronic breast cancer efficiently.