A Novel Machine Learning Method for Preference Identification: A Chess Case Study

This book introduces readers to a computational approach that learns, yet is unlike traditional machine learning. It does not rely on an artificial neural network (ANN) or large amounts of user data collected from online communities. Instead, the method uses collections of 'liked' and 'disliked' objects from a single user and from those is able to predict from a new collection which objects that user will most likely prefer or enjoy. Chess is used as the particular domain of investigation and experimentation (with chess problems as the specific 'objects'), even though the method is theoretically not limited to that domain. This book contains a detailed description of how the machine learning works and includes lots of readable programming code, database entries and many of the actual samples used in the experiments. Ideal for researchers and practitioners of machine learning, the material is also accessible to enthusiasts of both AI and chess.