Identification and Diagnosis of Breast Cancer At Different Stages By Different Machine Learning Algorithms On The Coimbra Dataset

Abstract

Cancer is the most deadly disease in the world. Breast cancer is the second-most common disease in women worldwide. It is the most common cancer globally among women. Annually, 12.5% of all new cancer cases worldwide Globally, 2.26 million breast cancers were discovered, and 685,000 women died from this disease. Early diagnosis of breast cancer is more difficult in developing countries than in developed countries. Using technology, if it is possible to detect cancer early and treat it on time, then many women can be cured and their lives can be saved. Early detection also leads to an increased survival rate for patients who receive clinical therapy before reaching later stages. It includes a number of risk factors, such as modifiable and non-modifiable ones. A recent survey discovered that for women above 50 years of age, the chance of getting breast cancer is about 80%. Machine learning algorithms are playing a major role in diagnosing liver cancer in its early stages and helping doctors make prompt decisions. A number of machine learning models have been executed in which the model gave better performance in terms of accuracy, and other parameters such as precision, recall, etc. are used to predict early. In this research work, the latest dataset, Coimbra, belongs to UCI machinery. It has nine features (age, BMI, glucose, insulin, HOMA, leptin, adiponectin, Resistin, MCP.1) and one classification attribute, which has values 1 and 2. 1 belongs to benign, and 2 belongs to malignant. Based on that, the supervised machine learning algorithm was applied. The WEKA tool is used to analyze the dataset. A number of algorithms are applied, such as Bayes net, multilayer perceptron, IBK, random committee, random tree, etc. More of them gave better results, and that model was chosen as the key model for breast cancer analysis.