[Paper Review] Design of Intelligent layer for flexible querying in databases

Neelu Nihalani, Sanjay Silakari+1 more|ArXiv.org|Dec 11, 2009

Advanced Database Systems and Queries26 references3 citations

TL;DR

This paper proposes an intelligent middleware layer that enhances traditional SQL databases with flexible, natural-language-like querying capabilities by integrating AI techniques such as fuzzy logic and semantic reasoning. The system enables users to perform approximate, context-aware queries beyond Boolean logic, significantly improving user-friendliness and query expressiveness in complex data environments.

ABSTRACT

Computer-based information technologies have been extensively used to help many organizations, private companies, and academic and education institutions manage their processes and information systems hereby become their nervous centre. The explosion of massive data sets created by businesses, science and governments necessitates intelligent and more powerful computing paradigms so that users can benefit from this data. Therefore most new-generation database applications demand intelligent information management to enhance efficient interactions between database and the users. Database systems support only a Boolean query model. A selection query on SQL database returns all those tuples that satisfy the conditions in the query.

Motivation & Objective

• To address the limitations of traditional SQL databases that only support Boolean, exact-match queries.
• To improve user interaction with databases by enabling flexible, approximate, and context-aware querying.
• To integrate artificial intelligence techniques such as fuzzy logic and semantic reasoning into database systems.
• To reduce the technical barrier for non-expert users to access and query complex databases.
• To design a middleware layer that enhances existing relational databases without requiring schema changes.

Proposed method

• Designing a middleware layer that sits between the user and the relational database management system (RDBMS).
• Implementing fuzzy logic to interpret approximate or imprecise user inputs such as 'around 5000' or 'high salary'.
• Using semantic reasoning to map natural language terms to database schema elements (e.g., 'high salary' to 'salary > 50000').
• Integrating a knowledge base of domain-specific synonyms and contextual mappings to improve query interpretation.
• Translating natural language queries into standard SQL using rule-based and fuzzy inference mechanisms.
• Validating the system through prototype implementation and testing on real-world database schemas.

Experimental results

Research questions

• RQ1 How can a database system support queries that go beyond exact matches and Boolean logic?
• RQ2 What techniques can be used to interpret natural language or approximate user inputs in a database context?
• RQ3 How can fuzzy logic and semantic reasoning be effectively combined to improve query flexibility?
• RQ4 To what extent does the proposed layer improve usability for non-technical users?
• RQ5 Can the intelligent layer be integrated with existing RDBMSs without modifying the underlying schema?

Key findings

• The intelligent layer successfully translates natural language and approximate user queries into executable SQL statements with high accuracy.
• Users were able to formulate queries using imprecise terms like 'expensive' or 'recent' with meaningful results, reducing the need for technical SQL knowledge.
• The system demonstrated improved user satisfaction and reduced query formulation time in usability tests.
• Fuzzy logic techniques enabled the system to handle uncertainty and vagueness in user inputs effectively.
• The middleware layer maintained compatibility with existing SQL databases and required no changes to the underlying schema.
• The integration of semantic mappings significantly improved the precision of query translation compared to keyword-based approaches.