Distance Measures Insights For Breast Cancer Analysis Using K-NN Algorithm
Dr.S. Bharathi, Krithika. L · Journal Of Advanced Zoology · 2024
Breast cancer stands as a prevalent disease among women, ranking high in terms of frequency. It is treatable if caught early enough. The greatest technique for predicting breast cancer is what this paper seeks to deliver. Mammograms can detect abnormal growths, although they are not always 100% accurate in identifying breast cancer. This article provides a superior way of prediction without biopsy, as it is currently not possible to confirm the presence of breast cancer without a biopsy. This study proposes the k-Nearest Neighbor (k-NN) technique, which is commonly used in machine learning for regression and classification. This study requires a number of steps, such as importing the dataset, pre-processing the data, and choosing the characteristics that need to be classified. The k-NN method additionally employed a variety of distance metrics to distinguish between benign and malignant tumours. Additionally, to demonstrate the effectiveness of the suggested strategy, the produced anser is contrasted with other outcomes. With improved distance measurements made possible by the k-NN algorithm, the study's findings advance our understanding of breast cancer prediction. Topsoe, Lorentzian distance, and Averageapproaches produced the most reliable overall results. Comparisons are made between the outcomes and established techniques such as the Euclidean, Clark, and Bray-Curtis distances.