Prediction of Reaction Force on External Indenter in Cell Injection Experiment Using Support Vector Machine Technique

Evaluation of the reaction force on a tool which is used for exertion of force on biomaterials such as biological cells or soft tissues has applications in virtual reality based medical simulators or haptic tools. In this study, two least square based support vector machine (SVM) models have been constructed to predict the indentation or reaction force on mouse oocyte and embryo cells in cell injection experiment. Inputs of these two models are geometrical parameters of indented cell, namely dimple radius (a), dimple depth (w) and radius of the semicircular curve (R). Experimental data for calibration and prediction of the models have been captured from literatures. The performance of the models has been evaluated using root mean square error (RMSE), correlation coefficient (r), relative error of prediction (REP), Nash-sutcliffe coefficient of efficiency (E_f) and accuracy factor (A_f). Comparison of the prediction results of the SVM models with experimental datapoints shows that the proposed SVM models have the potential to be used for force prediction applications.