Group II - Final Project Report (1) (1)

.docx

School

University Canada West *

*We aren’t endorsed by this school

Course

650

Subject

Business

Date

May 13, 2024

Type

docx

Pages

22

Uploaded by MateMoonYak12 on coursehero.com

1 Group II :- Final Project Report Olaoluwa Opaola – 2206218 Kunal Mittal – 2231823 Parveen Kataria – 2226061 Pamela Hernandez – 2242246 Malvika Makwana – 2303922 Arjunkrishna – 2305298 Anureet Kaur Dhaliwal – 2243237 Bhawana Sharma Koirala - 2311669 University Canada West BUSI 650-49 Business Analytics Sana Ramzan 26 th March 2024
2 Introduction The paper explains the regression analysis's outcomes on the financial data collected from both Fraud and Non-Fraud Companies and the narration of the key financial indicators descriptive analysis. Regression analysis is the tool to find out and describe the relationships inside financial indicators and the descriptive analysis serves the purpose of understanding the distribution and dispersion of these parameters. This analysis is performed by using predictive modelling techniques that are used in fraud detection, namely: decision tree, multiple regression analysis, clustering analysis, logistic regression, and neural networks. All the models are evaluated using different metrics and one of them is the accuracy, precision, recall, support, and F1, (Yacouby, R., & Axman, D., 2020). The main research question will be to figure out whether different analytics techniques can be somehow used to detect fraudulent activities within the data set. In order to answer these questions, I will conduct a number of sub-sections, for example: studying the descriptive statistics of the dataset, using linear regression models, creating visualizations using Tableau, and implementing predictive analytics using machine learning.
3 Brief Description of Techniques Descriptive Statistics: Descriptive analysis refers to the summarized information about the data, which includes the measure of central tendency as well as the dispersion measure, and the distribution, which is helpful to understand the characteristics of the data. Linear Regression: A basic linear regression model is applied to estimate the connection between dependent and independent variables. In the light of this, the learning will help to know how factors may contribute the occurrence of fraudulent practices. Tableau Dashboard: Tableau has the capability to create an interactive and informative dashboard with the aim of delving deep into data visually and uncovering hidden trends or patterns that would otherwise have remained untapped. Predictive Analytics (Machine Learning): Machine-learning algorithms like decision tree, logistic regression and neural networks can be used to identify fraudulent behavior on the basis of the patterns that are available from previous data. Analytical Reports Descriptive Report and Linear Regression: This section will give a brief descriptive report of the dataset’s characteristics, and then it will contain a linear regression analysis on the next five to study the connection between the variables. The database will be completed on Excel, containing the main findings from descriptive analysis and regression. The statistical analysis revealed various differences between the groups and made outliers or extreme values in many ratios and performance indicators
4 visible, (Yacouby, R., & Axman, D., 2020). The tail events often exhibited leptokurtic behavior, spanning equally to both sides, and were either positively or negatively skewed, based on the direction of the tail events. However, the regression result was the highest among the dependent variables with the return on assets as the independent variable (M-Score, Current Ratio, Quick Ratio, Accounts Receivable Turnover, and Asset Turnover). REGRESSION The first multiple linear regression was created by selecting the data of Fraud companies, with the dependent variable of Net Profit Margin, and the independent variables M score, Debt to Equity, Quick Ratio, Accounts Receivable Turnover, and Return on Assets. The Multiple R (0.13) shows a positive weak relationship and R Square (0.01769) a weak relationship between Net Profit Margin and the independent variables. Meanwhile Adjusted R Square is negative (- 0.0216115) which is less than R Square, interpreted as invaluable. Regression Statistics Multiple R 0.1329709 R Square 0.01768126 Adjusted R Square -0.0216115 Standard Error 987.252668 Observations 131 Regarding the relationship between the independent variable of Net Profit Margin with M Score, Debt to Equity, Quick Ratio, Accounts Receivable turnover, and Return on Assets, since the P-
5 value is greater than the significance level of 0.05, the relationship with the net profit margin of the fraud companies confirms as statistically insignificant. P-value Intercept 0.9825640 9 MScore 0.8732479 7 Debt to Equity 0.9951125 9 Quick Ratio 0.1448416 Accounts Receivable Turnover 0.9989029 6 Return on Assets 0.8675988 5 The second multiple linear regression was generated by selecting the same dependent variable but only of the Non-Fraud companies, and the independent variables, Mscore, Debt to Equity, Quick Ratio, Accounts Receivable Turnover, and Return on Assets. The Multiple R (0.17) shows a positive weak relationship and R Square (0.031) a weak relationship between Net Profit Margin and the independent variables. Also, Adjusted R Square (-0.027) is invaluable as it is less value than R Square. Regression Statistics Multiple R 0.176581246
6 R Square 0.031180936 Adjusted R Square 0.027816981 Standard Error 92.05339107 Observations 1446 The results of the P-value of these variables is greater than 0.05, meaning again the relationship is statistically insignificant between the data of the Non-Fraud Companies. P-value Intercept 0.257399283 MScore 1.96937E-07 Debt to Equity 0.597260405 Quick Ratio 0.053579691 Accounts Receivable Turnover 3.11573E-05 Return on Assets 0.51247816 The third multiple linear regression was with the selection of Mscore data only of Fraud companies as the dependent variable and the independent Variables of Gross Profit Margin, Debt on Equity, Current Ratio, EBITDA, and Return on Equity. The Multiple R (0.024) shows a positive weak relationship and R Square (0.00059) a weak relationship between Net Profit Margin and the independent variables. Also, Adjusted R Square (-0.0393) is invaluable as it is less value than R Square. Regression
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help