An expert system to recommend fiction to the casual reader. Tentatively named "Fantastic Books and Where To Find Them" or "Book Guru".
An expert system is a form of AI software popular in the 1980's - in fact, it was one of the first successful forms of AI. An expert system is supposed to emulate the expertise of a human who has practiced their craft for tens of thousands of hours. It does this by having a set of rules and facts.
This expert system currently recommends teen and adult fiction using the Java based rule engine Jess and the CLIPS programming language. It has rules for asking questions and facts for all possible books that the engine can recommend. Jess doesn't sequentially go through code, but instead executes a rule whenever the left hand side of the rule is satisfied (basically like an if statement). It is unpredictable when a rule will be executed and at the same time possible for a rule to be executed multiple times. This particular system questions the user using the question rules until the recommend rule is fired. The recommend rule is fired when there is only one book left in the library. Whenever a question is answered books are retracted from the library until there is only one book in the library.
- Download Jess from https://www.jessrules.com/jess/download.shtml. This should give you a 30 day trial to the software. I am currently waiting for Jess to be open sourced which should happen sometime in 2019 hopefully.
- Save book_recs.clp in the examples/jess folder.
- Run Jess using jess.exe in bin. Command would be bin/jess.
- Run (batch "examples/jess/book_recs.clp") in Jess.
- Add more books to the library, preferably adult fiction and literature.
- Add length of book as a characteristic. Short, medium, long?
- Convert code into LISP or another rule engine based coding schema which doesn't have a 30 day trial.
- Fix the plot-driven vs character-driven question - people find it confusing.
- Get rid of compelling as a characteristic.
- Make it so that if the first question is given a book the system doesn't know, it saves that book into the database.
- Standardize which characteristics go with which appeal factor.
- Make the final decision print out what characteristics made it go to that decision - makes decision more interpretable.
- Add a book! Genres can be verified by looking at the Wikipedia page of the book.
- Update the characteristics of a book.
- Wikipedia for background information on expert systems
- Dr. Nelson for hosting an expert systems class
- Ms. Cranston & Ms. Vaughan for helping me determine categories of appeal factors in books