Analysis and prediction of stomach cancer using machine learning

Abstract

Cancer prediction and analysis systems offer aid in the management of patients and have been found to provide accurate forecasts for stage and survival prediction. This study presents a cancer prediction system developed using machine learning models and implemented with Streamlit. This system is capable of accurately predicting cancer stage onset along with chances of the patient’s onset of survival based on prior patient information. For predictive purposes, categories such as random forest and XGBoost were employed. The model achieved an effective accuracy of 85% for stage prediction and 97% for predictability of patients’ survival. This application includes a simple interface that healthcare professionals can employ to enter patient data and immediately make educated predictions. This paper illustrates the assistance these integrated systems provide clinicians and how they can ameliorate functional healthcare practices. In the future we are hopeful and aim towards further increasing the strength and efficiency of the system by enhancing the dataset used and additional predictive models.