Improving the Performance of Machine Learning Algorithms Using Conceptual Models - A Case Study of Auto Insurance

Aug 3, 2026

Upload by Ophelia

Domain knowledge integration is proposed to mitigate the limitations of purely data-driven machine learning, where fixed data structures can miss context-specific insights. This thesis evaluates how adding domain knowledge during preprocessing and feature engineering affects decision tree and Support Vector Machine (SVM) models in auto insurance. Models trained on an original dataset (A) are compared with models trained on a revised dataset enhanced with domain-specific guidance (Revised A). Results show substantial accuracy and reliability gains, especially when domain knowledge refines features related to customer profitability, while smaller improvements appear when domain influence is limited. Overall, benefits are meaningful but context-dependent.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages67

Reading Time2 h 49 min

AccessFree to download

1. Introduction

2. Background

2.1 Introduction

2.2 Key Concepts and Context

2.2.1 Machine Learning

2.2.2 Domain Knowledge

2.2.3 Conceptual Model

2.2.4 Customer Profitability

2.2.5 Car Insurance

2.3 Theoretical Background

2.3.1 Theoretical Framework of Machine Learning (ML)

2.3.2 ML predictive analysis

2.3.3 Improving ML Predictive Analysis Performance

2.3.4 ML and Domain Knowledge

2.3.5 Conceptual Models in Machine Learning

2.4 Research Background

2.4.1 Benefits and Challenges of ER Modeling in ML

2.4.2 Recent Advancements in ER Modeling

2.4.3 Key Benefits of EER Models in ML

2.4.4 Applications and Case Studies

2.4.5 Conceptual Models for Machine Learning (CMML)

2.4.6 Trends in Integration

2.4.7 Gaps in Current Research

2.4.8 Emerging Themes

2.4.9 Summary of Literature Review

Improving the Performance of Machine Learning Algorithms Using Conceptual Models: A Case Study of Auto Insurance by © Maedeh Moosavi (Thesis) submitted to the School of Graduate Studies in partial fulfillment of the requirements for the degree of MSc. Management, Faculty of Business Administration Memorial University of Newfoundland Feb 2025 St. John’s Newfoundland and Labrador Abstract The integration of domain knowledge into machine learning models has been proposed as a means to address the limitations of purely data-driven approaches. Traditional machine learning techniques often rely on pre-defined, fixed data structures, which can overlook valuable context-specific insights that domain knowledge provides. This study investigates the impact of incorporating domain knowledge into the preprocessing and feature engineering stages of machine learning models, specifically focusing on decision tree algorithms and Support Vector Machines (SVM) within the insurance sector. To evaluate the effectiveness of this integration, this study compares the performance of models trained on a pre-defined dataset (A) with models trained on the same dataset after it was enhanced with domain-specific knowledge (Revised A) . The results demonstrate that the integration of domain-specific guidelines into the feature engineering process significantly improved the accuracy and reliability of the predictive models, particularly in complex scenarios such as predicting customer profitability. In scenarios where domain knowledge played a crucial role in refining features that capture relationships within the insurance data, the enhanced models outperformed the original ones. Conversely, for tasks where the domain knowledge had less influence, the performance improvement was marginal. These findings suggest that integrating domain knowledge into machine learning processes can provide a meaningful boost in model effectiveness, but the benefits are context-dependent. Keywords: Machine learning, domain knowledge, feature engineering, decision trees, SVM, insurance industry, predictive modeling. Acknowledgments I would like to express my deepest gratitude to my supervisor, Dr. Jeffrey Parsons, for his unwavering support, insightful guidance, and encouragement throughout this journey. His expertise and advice were invaluable to the completion of this thesis. Table of contents 1. Introduction……............................................................................................................. 1 2. Background…….............................................................................................................2 2.1 Introduction………………………....................................................................................................................2 2.2 Key Concepts and Context…………........................................................................................................................2 2.2.1Machine Learning……………………...................................................................................................2 2.2.2Domain Knowledge………………….....................................................................................................2 2.2.3Conceptual Model…………………........................................................................................................ 3 2.2.4Customer Profitability………………….................................................................................................. 3 2.2.5Car Insurance…………………..............................................................................................................4 2.3. Theoretical Background…………….................................................................................................................4 2.3.1. Theoretical Framework of Machine Learning (ML) .............................................................................4 2.3.2. ML predictive analysis................................

FAQs

What problem does the thesis address in machine learning modeling?

It addresses limitations of purely data-driven approaches that rely on fixed data structures and can overlook valuable context-specific insights provided by domain knowledge.

Which models and stages are investigated for domain knowledge integration?

The study focuses on decision tree algorithms and Support Vector Machines (SVM), examining the effect of domain knowledge added during preprocessing and feature engineering.

How are the results evaluated in the case study?

Performance is compared between models trained on the original dataset (A) and the same dataset enhanced with domain-specific knowledge (Revised A), with gains particularly noted for predicting customer profitability.

Aug 3, 2026

Upload by Ophelia