In this study, we define the concept of fuzzy parameterized fuzzy soft matrices (fpfs-matrices) and present some of their basic properties. By using fpfs-matrices, we then suggest a new algorithm, i.e. Prevalence Effect Method (PEM), and apply this method to a performance-based value assignment, so that we can order noise removal filters regarding performance. The results show that PEM has potential for several areas, such as machine learning and image processing. Finally, we discuss fpfs-matrices and PEM for further research.