Article Preview
TopIntroduction
Nature is obviously a perfect example for optimization since it inevitably finds an optimal way to solve complex relationships among living creatures ranging from microorganisms to fully fledged humans including ecosystem balancing, diversity maintenance, climate adaptation and so on. Although the concept behind nature is simple, the outcomes are remarkable and unbelievable. Nature is also the best teacher and its policies and capabilities are huge that researchers are even mimicking the behavior of nature in their technology (Binitha & Sathya, 2012). Numerous optimization algorithms are henceforth developed from the behavior of some animals or insects in nature like ant colonies, bees swarm, and so on. This is because the biological activities of birds and animals are responsible for specific roles both individually and as a group, to achieve a specific task in their daily routine or lifetime. Accordingly, they have received the attention of data analysers to resolve numerous difficulties in science and engineering sector (Darwish, 2018). Bio-inspired optimization has been emerging as a new era in computation that supports huge variety of applications covering almost all fields like robotics, security, networks, medicine, data mining, controllers, power devices, production units and so on. Moreover, the improvements made in these algorithms are fully inspired from the entire behavior of natural activities. These inspirations experienced from nature possess the capability to define and resolve complex processes with inherently effective initial features and strategies in addition to little or no knowledge about search domain.
The use of bio-inspired optimization in the field of data mining is increasing day-by-day. Data mining is termed as the process of extracting valuable information from a database or dataset and developing a practice that can be effectively applied to identify the structure of data. It is one of the fastest growing research areas in the information field nowadays because of the widespread accessibility of excess amount of data and the importance to convert those data to valuable information (Han et al., 2011). Data mining methods have been recognized as one of the most suitable means to extract useful information from huge databases in knowledge discovery data process. The extracted data can be utilized for wide range of applications including predictive modelling, operation optimization, fault detection and diagnosis, classification and so on (Zhao et al., 2020). Of these, classification is commonly used in many fields to classify a large dataset into predefined classes, labelled by different attributes (Fayyad et al., 1996). For example, classification is applied in medical image diagnosis to predict the status of a patient whether he/she is ill or normal. However, classification mainly uses labelled data for developing its model to train the classifier. That is, every object of the labelled dataset taken for training has been assigned to exactly one individual class, which refers to a particular attribute termed as class attribute. The models that are designed based on this consideration can be applied to classify new datasets (without any class label attributes), extract relevant piece of information, predict future trends, valid patterns, etc.