[Paper Review] A Framework for Business Intelligence Application using Ontological Classification

Martin Aruldoss, D. Maladhy+1 more|arXiv (Cornell University)|Sep 6, 2011

Big Data and Business IntelligenceBusiness, Management and Accounting6 references19 citations

TL;DR

This paper proposes a Business Intelligence framework that leverages ontological classification and decision tree mining to extract and structure competitive intelligence from web data. By modeling domain-specific relationships through an ontology and applying decision trees to discover patterns, the framework enables more accurate and actionable business insights from unstructured web content.

ABSTRACT

Every business needs knowledge about their competitors to survive better. One of the information repositories is web. Retrieving Specific information from the web is challenging. An Ontological model is developed to capture specific information by using web semantics. From the Ontology model, the relations between the data are mined using decision tree. From all these a new framework is developed for Business Intelligence.

Motivation & Objective

• To address the challenge of extracting specific, structured competitive intelligence from unstructured web data.
• To develop an ontological model that captures domain-specific relationships and entities relevant to business intelligence.
• To integrate decision tree algorithms for mining meaningful patterns from the structured data in the ontology.
• To create a comprehensive, reusable framework that enhances business intelligence through semantic web technologies and data mining.

Proposed method

• An ontological model is designed to represent business entities, their attributes, and inter-relationships using web semantics.
• The ontology is populated with data extracted from relevant web sources using information extraction techniques.
• Decision tree algorithms are applied to the annotated data to discover predictive relationships and classify business intelligence patterns.
• The framework integrates the ontology and decision tree outputs into a unified system for business intelligence applications.

Experimental results

Research questions

• RQ1 How can ontological classification improve the accuracy and structure of competitive intelligence extraction from web data?
• RQ2 What role do semantic relationships in an ontology play in enhancing decision-making in business intelligence?
• RQ3 How effective are decision trees in identifying meaningful patterns from ontology-structured business data?
• RQ4 Can a combined ontology and data mining approach yield a scalable and reusable framework for business intelligence?

Key findings

• The proposed framework successfully structures unstructured web data into a semantically rich ontology, enabling more precise information retrieval.
• The integration of decision trees with the ontology enhances pattern recognition, leading to more actionable business insights.
• The framework demonstrates improved data organization and relationship discovery compared to traditional keyword-based approaches.
• The use of ontological classification reduces ambiguity in data interpretation, increasing the reliability of extracted business intelligence.