A method of identifying cell suspension concentration based on bioimpedance spectroscopy

Abstract

Based on bioimpedance spectroscopy technology, a method of automatically identifying the cell suspension concentration is proposed. This method combines multiple linear regression algorithm and bioimpedance spectroscopy technology, which can identify the concentration of cell suspension quickly and accurately. Firstly, a strategy of random distribution of cell locations is proposed to simulate the true existence of cells. Secondly, 2400 groups of normal, cancerous and mixed cell models with different concentrations are generated by numerical simulation and their bioimpedance spectroscopy data are calculated.Thirdly, the multiple linear regression algorithm (MLR), support vector machine (SVM), and gradient boosting regression algorithm (GBR) are used to identify the concentration of cancerous cells. The simulation results show that the MLR is the best regression model for cell suspension concentration identification and its average goodness of fit and mean square error are 0.9997 and 0.0008respectively. Finally, the MLR is applied to the identification of red blood cell suspensions with different concentrations, the experimental results show that the average goodness of fit and mean square error are 0.9998 and 0.0079, respectively, indicating that this method has a greater ability to identify cell suspension concentrations.