Electrochemical biosensors are leading the way in the development of point-of-care devices. Many of these biosensors predominantly use electrochemical impedance spectroscopy for their detection mechanism. The standard method for creating calibration curves in impedimetric biosensors is to fit the impedance spectrum to an equivalent circuit model(ECM). The calibration curve is then created using the circuit parameter that shows the maximum variance. This standard approach has multiple limitations. First, in a complicated system, the identification of an ECM that correctly describes the system is not a trivial task as it might involve complicated processes. Second, in many cases, multiple ECMs could fit a given impedance spectrum which could introduce errors or biases in the analysis. To overcome these limitations, in this work, we use machine learning (ML) algorithms to develop calibration curves for impedimetric sensors without the use of equivalent circuit models or by manually picking points from the impedance spectrum. For this, raw impedance spectrum corresponding to different impedimetric sensors were obtained from the literature and a labeled dataset was created for training the ML model. Principal component analysis was done to extract the features that show maximum variance. These features were then used to train multiple machine-learning models to create a calibration curve. The errors of the different models are compared to identify the best-performing ML model. Finally, a comparison was made between the calibration curve obtained from the conventional approach and that obtained from the ML model. This comparison shows that errors were lower for predictions made using the calibration curve obtained from the ML model.