Every day the number of datasets is increasing, and the advancement also leads to different research algorithms to extract meaningful information about personal interests. The problem of frequent itemsets from analysing different periodical itemsets from the large chunks of data has always been a keen research area for data analysts. Therefore, the most valuable fields of data mining are computational high-utility itemset mining i.e., used to obtained valuable knowledge from a homogeneous mixture of data. In the past years, various algorithms and works are proposed by different researchers to solve the issues of high-utility or frequent itemsets mining in transactional databases; however, an essential limitation of the algorithms is static nature [14]. In real-life applications, datasets are often dynamic and change according to different parameters such as price, weight, quality etc. Therefore, a transformation of the existing database and redefinition of parameters is required for gaining maximum profits from the high-utility itemsets mining. In this research, a novel approach is designed and developed i.e., a computational intelligence approach based on high-utility itemset mining algorithm for efficient analysis of periodical itemsets from the transactional database [19]. The proposed MFHUIM algorithm is further experimented on retail BMS database using Java-based environment and results of both the algorithms based on runtime, memory usage and scalability are detailed in the research work. MFHUIM outperforms EFIM algorithm in extraction of the accurate high utility itemsets in a large set of databases.