Construction of an Intelligent Book Recommendation System for Libraries Under Collaborative Filtering Algorithms

With the development of information and network technology, the scale of data on the Internet is unprecedented. Big data and artificial intelligence have generated widespread intersection and integration in various fields. In social platforms, e-commerce, and online services, personalized recommendation has gradually become a basic function, and the era of intelligent recommendation has arrived. Searching for the desired bibliography from thousands of books is a very labor-intensive task. Collaborative filtering is a machine learning based human portal algorithm that intelligently recommends the similarity of different types of books, greatly saving users time and improving the efficiency of book borrowing. This article explores a personalized recommendation method for a library intelligent book recommendation system based on collaborative filtering algorithms (CFA), taking reader reading data and library bibliography recommendation strategies as research objects. The borrowing records generated in the library reflect the reading characteristics of readers, with readers as the basic unit. There is a significant bias in the preferences of each reader in the library, so centralized processing of records is not accurate for recommendation systems. Therefore, incorporating a comprehensive consideration of the attributes of bibliographic items and reader preference elements into the algorithm can provide more accurate book recommendations. The experimental results indicate that the CFA optimized based on K-means algorithm has a shorter response time in recommending books to readers compared to traditional CFA. The optimized algorithm has an average response time of 0.28 seconds, which is 0.18 seconds less than traditional CFA.