Pattern-based market basket analysis is an interesting area of research in recent times. Pattern explores hidden knowledge from the database that helps in smart decision-making for business promotion. In this context, periodic frequent pattern mining has a great significance in knowledge discovery. However, the traditional approaches follow pattern pruning with support and periodicity threshold only, which often compromises the quality of the patterns in the context of correlation. Some of the approaches consider correlation metrics in pattern generation to solve the problem, but the selection of the right correlation measure and specification of the correlation threshold remains a problem. This paper introduces a new pruning mechanism called dissociative pruning to extract periodic correlated frequent patterns from the database. The proposed method is free from the selection of correlation measures as well as the specification of the correlation threshold. Dissociation between the itemsets is the backbone of the dissociative pruning model. Experimental outcomes from real databases depict the acceptability of the proposed approach in various contexts, such as correlation, pattern explosion, runtime, and scalability.