Article Preview
TopIntroduction
Nowadays, the recommendation system plays a crucial role in the e-commerce platform. Recommender system has its role in delivering more precise and reliable data to particular users (Bobadilla et al., 2013; Ortega et al., 2016). Recommender system will gather the relevant data and opinion for the user or set of the group. For example, if the new user entered for registering in the Amazon website, then recommendation engine attempt to examine the action and pattern of the user. If the recommendation system is unable to offer the required information, it leads to cold start problem (Son, 2016; Ji & Shen, 2016; Zhao et al., 2016). In this scenario, system unaware of the new user’s opinion or likes, but user spent time on a specific website; then the system will collect information about user preference about the items. Tourism, online shopping, entertainment, web intelligence are some business applications that seen in a recommender system, but still, it moves toward the development stage. The primary recommender system performs in the different filtering process, namely collaborative filtering (CF), context-based (CB) and hybrid. CF is the most preferably applied in RS (Hernando et al., 2016; DaSilva et al., 2016; Liang et al., 2016; Yang et al., 2016; Zhou, 2016; Maheshwari & Prasanna, 2016).
In collaborative filtering (CF) based system, the recommendation will offer an opinion by the rating given by users. For example, when users provide a rating for an e-commerce platform for various items like laptop, dress, then the recommender system will give preference to a specific product to the user. In context-based (CB) filtering, it mainly concentrates on the decision. It may rely on the choice presented by previous users (Xu & Yin, 2015; Puglisi et al., 2015). For example, if a user purchases a laptop, then RS will provide suggest only similar usage of laptop items because RS will find-out action of this individual person. In hybrid based recommendation system, it combined with CF, CB with various machine learning techniques and algorithms (De Pessemier et al., 2017; Moradi & Gholampour, 2016; Capdevila et al., 2016). To acquire an optimized recommend result is always a challenging task, mainly in multiple domains like computational intelligence, data mining, and machine learning. The bio-inspired algorithms are most predominant and competent in data optimization. Ant colony optimization (ACO), Particle swarm optimization (PSO), Bacteria foraging optimization (BFO), Genetic algorithm (GA) are some optimization technique to get accurate results in specific domains like image processing, data mining and bioinformatics (Yang, 2010). These algorithms can apply in a recommender system as an expert recommendation. We have used the fireflies algorithm; it is based on a meta-heuristic approach. The primary objective of fireflies algorithm is to the performed signal system to pull other another firefly. Firefly algorithm provides efficiency for a specific problem, and it needs only a fewer number of iterations (Yang, 2010; Lones, 2014; Weyland, 2015). Then apply FCM for classification of users in the dataset by using the similarity of user ratings (Bezdex et al., 1984; Cannon et al., 1986; Havens et al., 2012; Sengupta et al.,2018). Datasets are categories as training data and test data into the ratio of 80:20. The main effort of this research aspect is: