The challenge of identifying the location and degree of structural damage from measured modular test data is discussed in this work. Reverse engineering is used in this procedure to find structural element faults. Direct techniques, such as the dynamics method, are used to obtain vibration test data. However, traditional approaches are expensive and time-consuming. Sugeno–Takagi FIS has thus been employed in this instance to train the data and discover a global solution for crack location prediction. However, it has been noted that the Sugeno–Takagi FIS is insufficient for training the dataset. Most of the time, mistakes and uncertainty are present in the data that was gathered to create the data set. Even while applying machine learning and artificial intelligence techniques has many benefits, they are nonetheless limited by the lack of a data set. Errors are added to the data set as the data is being collected. Therefore, a cleaning procedure that reduces mistake must exist. In this study, the Sugeno–Takagi FIS was subjected to the boot strapping method, and the outcomes are contrasted.