In this paper, the performances of two data-driven dynamic analysis methods, namely exDMD and LANDO, are compared for data-driven modeling of a dynamic system related to cancer. The data used in this paper consists of 1000 samples with 33 features. The exDMD method, with a linear approach, performs well in analyzing almost linear systems, but is limited in dealing with nonlinear behaviors. In contrast, LANDO, using kernels and sparse dictionaries, effectively identifies linear and nonlinear components of the system. In this study, after a conceptual and mathematical explanation of both methods, the results of their implementation on data obtained from the mathematical model of cancer and double pendulum in the MATLAB environment, and by applying different levels of noise to the data, are investigated. Comparison of the outputs shows that the LANDO method has higher accuracy in modeling nonlinear behaviors. Even the prediction made using the linear operator obtained through the LANDO method is more accurate than exDMD. These results can help improve the analysis and prediction of the dynamics of diseases such as cancer. In the results section, to show the superiority of the LANDO method over the exDMD method, the performance of the two methods on the double pendulum data is also compared, and the results confirm the better performance of LANDO over exDMD.