Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 

Bookshelf

This application is based on ontology schema and productions.

Ontology is the description of the world based on entities, attributes and relationships between entities. Ontology is often presented as ordered structure of concepts of an specific shpere.

Attribute is property of an entity.

Production - rules stored in knowledge base. While system parses user query it searchers the most relevant rule for the answer generation.

In general case rules can be persented as context-free grammar. In our case productions are patterns for synonyms and sub-categories.

How would it work?

  1. User types the query
  2. System tokenize this query and removes stopwords
  3. For any word of the query:
    • searcher find matches in all fields of the book
    • otherwise, system calls the knowledge base and find answers by the query context

How to lmprove the current algorithm?

  1. Add new books in the JSON file
  2. Write rules either for herous if book is feature or for the sub-category