Breast Cancer Prediction Based Clinical Symptoms Using K-Means

One of the highest contributors to death in 2022 in Indonesia is breast cancer. This condition is a disease of breast cells that develop abnormally and out of control. If it is not treated immediately, it is feared that it will spread to other parts of the body, including the bones, liver and lungs. The aim of this paper is to predict based on clinical symptoms of breast cancer using K-Means. It is hoped that this approach will be a simple approach but has a good level of accuracy. The data considered in this study were clinical data consisting of BMI (kg/m2), age (years), insulin (
$\mu\mathrm{U}$
/mL), glucose (mg/dL), Homeostatis (HOMA), adiponectin (
$\mu\mathrm{g}$
/mL), leptin (ng/mL), MCP-1 (pg/dL), and resistin (ng/mL). The stages of this research consisted of data analysis, data normalization, modeling based on clinical data using K-Means, and further testing using both data testing and data learning. Tests using training data obtained accuracy value 78%, a precision value of 0.84 and sensitivity value 0.84. The test results using data testing obtained accuracy value 75.76%, a precision value of 0.76, and a specificity value of 0.710.