부산대학교 도서관

Each day, a vast amount of data is generated by a range of medical sources (actuators, devices, and sensors). Medical researchers and physicians can discover a wealth of valuable medical information by examining this medical data. This information can be utilized to detect diseases at an early stage and to make critical treatment decisions. However, in the era of big data, the machine learning community faces new hurdles when applying classification algorithms to real-world scenarios due to the data's variety, volume, velocity, and validity. The primary goal of this research is to automate the learning procedure (feature extraction, parameters tuning, and optimization) of the existing classification algorithms such as support vector machine (SVM) and random forest (RF). For this proposed method, Chi-square feature analysis is utilized to automate the feature selection process, and principal component analysis is being used to reduce the dimensionality of large datasets. Further, to improve the parameter tuning of machine learning algorithms, an improved Imperialist Competitive Algorithm (ICA) is employed in this article. Finally, S VM and RF classification models are created and compared to find the ideal combinations for big data analysis.