Recommender systems are software applications that offer customers suggestions for products based on their previous purchases or item evaluations. Collaborative filtering (CF), among the most effective recommendation techniques, has garnered substantial attention and practical implementation across academic and business settings. A prime strategy for constructing recommender systems is through collaborative filtering (CF), a method that extrapolates forecasts or recommendations about future user preferences by analyzing the collective preferences of a user group, all without delving into the specifics of the content itself. In collaborative filtering recommender systems, user inclinations are conveyed as ratings assigned to items, with each new rating augmenting the system’s knowledge and influencing the accuracy of its suggestions. Generally, as customers contribute more ratings, the impact of the suggestions grows. Nevertheless, the significance of each individual rating may differ significantly; various ratings can supply diverse quantities and types of information about a user’s preferences. This article delves into the recent innovations suggested for collaborative filtering and the most up-to-date uses within the realm of recommendation systems. Furthermore, it conducts a hypothetical case study involving a recommendation system that employs diverse collaborative filtering algorithm techniques. The aim is to distinguish the appropriate quality metrics and indicators for evaluating both categories of collaborative filtering algorithms: memory-based and model-based.