Rule extraction using fuzzy C-means clustering and an extreme learning machine

Nowadays by the increasing volume of data in databases, knowledge discovery has received due attention for accurate and fast processing of data. Rule extraction is known as an efficient way for knowledge discovery by using data mining and machine learning techniques. The extracted rules can be employed in rule learning using machine learning methods in order to be used in decision-making systems. Rule learning methods describe the dataset in the form of decision rules. In this paper, a new method for extracting rules from raw data is introduced based on fuzzy c-means clustering algorithm, and then, learning the rules using the extreme learning machine (ELM). The method first divides the dataset into partitions based on type of classes. Then, the fuzzy c-means algorithm is used to produce clusters. Each cluster is depicted as an if-then rule and considered the input of the ELM machine. The ELM machine is used for learning the rules for the final goal of decision making. The proposed method is evaluated and compared with two other rule learning methods. The results revealed the preference and applicability of the proposed method.