Although experimental modal analysis (EMA) is a common tool in structural dynamics, it is still mostly performed by experts. To make it available to every user, many challenges must be addressed throughout the process of planning the EMA, conducting the measurements, extracting the modal parameters, and interpreting the results.This paper focuses on the process step of parameter extraction in which the modal parameters (eigenfrequency, damping, and mode shape) are estimated to fit the measured data. The Least Squares Complex Frequency Method (LSCF-Method) is a robust and efficient method widely used to perform this task. However, it requires the user to interact and define input-parameters to gain reliable results.Within the scope of this work a neural network, which assists the user to parametrize the LSCF-Method, was developed, trained, and tested.