Comparison the effects of KNN and linear regression models in lung cancer prediction

Abstract

Lung cancer has a range of major factors like smoking, yellow gingers, anxiety, etc. now, the problem of this research is that prediction for lung cancer. Prediction for lung cancer is a complex problem that is not suitable for human prediction. This research using a dataset was from Kaggle. There are 16 rows and 309 columns. To determine the k nearest neighbors (KNN) algorithm and linear regression algorithm, which one is better for prediction for lung cancer, and which coefficient will be best effective. This research uses mixed method research. In this work, when the K of the KNN algorithm equals 7 or 2, the effectiveness of the KNN model is best, when the alpha of the linear regression algorithm equals 20, the effectiveness of the linear regression model is best. The KNN model is better than the linear regression model, though the difference is negligible. In the future, more emphasis can be placed on using a wider range of algorithms or using more extensive and generalized dataset, as well as assessing the efficiency of the algorithm on larger datasets.