Recommender systems handbook Book 3

This week I saw chapter 6. The book consists of 25 chapters.

From the point of view, this book provides a comprehensive introduction to the recommendation system, and also introduces some specificAlgorithm. There are some mathematical symbols in these formulas that I can't remember.

The following is a summary of the first six chapters:
Chapter 1: Introduction to the book;
Chapter 2: Data mining methods used in recommendation systems, divided into: Data Processing (similarity measurement, sampling, dimensionality reduction, and noise reduction), classification (specific algorithms include nearest neighbor, decision tree, rule-based classification, Bayesian classification, artificial neural network, and support vector machine), clustering analysis, and Association Rule Mining
Chapter 3: Content-based recommendation system: State of the Art and trends.
Chapter 4: Overview of neighbor-based recommendation Methods.
Chapter 5: Improvement in Collaborative Filtering;
Chapter 6: Develop a constraint-based recommender;

The following content is taken from Chapter 6:

Traditional recommendation methods (Content-Based Filtering and collaborative filtering) are very suitable for products such as books, movies, and news. However, it is not the best recommendation method in the automotive, computer, real estate, financial services, and other fields. For example, the number of real estate transactions is much smaller, and a product is not easy to collect a large number of user reviews. In addition, the user is not satisfied with the recommendation based on the product features of a few years ago.
Knowledge-based recommendation systems can solve such problems, and knowledge-based recommendation systems are not cold-started (new products cannot be recommended. Of course, knowledge acquisition is the bottleneck of such systems.