Many real world applications contain a decision making process which can be regarded as a pattern classification stage. Various pattern classification techniques have been introduced in the literature ranging from heuristic methods to intelligent soft computing techniques. In this paper, we focus on the latter and in particular on fuzzy rule-based classification algorithms.We show how an effective classifier employing fuzzy if-then rules can be generated from training data, and highlight how the introduction of class weights can be used for costsensitive classification. We also show how a training algorithm can be applied to tune the classification performance and how genetic algorithms can be used to extract a compact fuzzy rule base. We also give pointers to various applications where these methods have been employed successfully.