Predicting default probability in construction industry basing on Over-Sampling Technique for Grey SystemTheory

Bankruptcy Prediction has been a hotly-debated topic among many people in business area. The fact is that once the firm goes bankrupt, it will be disastrous to not only firm itself but also other stakeholders. Many available methods have been applied to predict the possibility of business collapse; almost all of them were based on financial ratio analysis. Grey System Theory, used in the previous thesis for predicting default probability of construction firms, has brought some feasible results, by relying on the 19 initial financial ratios. This study, with the aim of enhancing the Grey Theory application, employs Over-sampling technique before applying Grey System Theory. The results of this study are then compared with those of the prior research. Furthermore, replication and Synthetic Minority Over-sampling Technique (SMOTE), two over-sampling techniques are proposed to resolve the imbalance problem in data set. The results reveal that over-sampling techniques could improve the predicting performance of Grey System theory. Additionally, between these two kinds of over-sampling techniques, SMOTE surpasses Replication in terms of prediction capability.ACKNOWLEDGEMENT ii ABSTRACT iii TABLE OF CONTENTS iv LIST OF FIGURES vii LIST OF TABLES viii CHAPTER 1:INTRODUCTION 1 1.1 Background 1 1.2 Motivation and Problem statement 2 1.3 Research Objectives 3 1.4 Research Scope and Limitation 4 1.5 Thesis Structure 5 CHAPTER 2: LITERATURE REVIEW 7 2.1 Default Prediction Research 7 2.2 Default Pediction Research In the Construction Industry 12 2.3 Grey System Theory in Prediction Default Probability 15 2.4 Summary 18 CHAPTER 3:METHODOLOGY 19 3.1 Grey System Theory 19 3.1.1 Methods of Grey Numbers’ Generation Based on Average 20 3.1.2 Grey Incidence Analysis 22 3.2 Over-Sampling technique…………………………………27 3.2.1 Between-class Imbalance Problem in Date Set .......................................... 27 3.2.2 Replication ..................................................................................................... 28 3.3.3 Synthetic Minority Over-sampling Technique (SMOTE) ........................ 28 3.3 ROC Curve 31 3.3.1 Concept and Methodology of ROC Curve 31 3.3.2 Utilize ROC Curves to Validate the Model……………………………........... 33 3.3 Summary 34 CHAPTER 4:DATA COLLECTION 35 4.1 Data Colection 35 4.1.1 Source and Validity of Data 35 4.1.2 Principles of Collecting Data………………………………………………..... 35 4.1.3 Summary of the Input Data……………………………………………............ 36 4.2 Data Clasification 37 4.3 Financial Ratios Definition 39 4.4 Data analysis process 44 4.5 Summary 44 CHAPTER 5:DATA ANALYSIS AND RESULTS 46 5.1 Data Analysis 46 5.1.1 Example Analysis 46 5.2 Results 56 5.2.1Reasonable Data Consequence 56 5.2.2 Results of previous study (Le Quyen’s thesis).…………………..……….…... 57 5.2.3 Comparisons………………………………………………..……..…………..... 59 5.3 Summary 60 CHAPTER 6:CONCLUSIONS 61 REFERENCES 62 APPENDICES…………………………………………………………………………65 A.1 Data Collection of Construction Firms 65 A.2 Data Collection of 15 Firms for Example Mathematics 6