Abstract
The goal of this project, is to compare the AFT model and Cox PH model using the employee attrition data set. Survival analysis examines the desired outcome until the occurrence of the event. Although Cox PH together with AFT models have been widely utilized in survival time predictions, AFT models are least used in employee attrition. Therefore, the goal of this research is to conduct survival analysis on the employee attrition data set to narrow down on the specific factors that will benefit the employer using both models and the best method to use. Using R, the Accelerated Failure Time model gave favourable outcome compared to the Cox PH model. The main factors that have a significant impact on the survival attrition include, the job role(Research Scientist, Sales Executive), home to job distance, work life balance, level of satisfaction in job and nature of travel. Furthermore, the Generalized Gamma AFT model offers the most outstanding fit for the observed data. The research will serve as a focal point for surviving analysis models in predicting employee attrition, enlightenment in the analysis and deepen the context of survival analysis.
Table of Contents
Abstract ii
Declaration and Approval iii
Dedication iv
Chapter One
Introduction
1.1 Background Information 1
1.3 Objectives 2
1.4 Significance of the Study 2
1.5 Scope of the Study 3
Chapter Two
Literature Review
2.1 Staff Attrition 4
2.2 Approaches to Employee Attrition 5
2.3 Non-Parametric and Parametric Models Approaches 5
3 Methodology 8
3.3 Variables of the Predictor 8
3.4 Study Design 8
3.5 Methods
3.5.1 Survival Analysis 8
3.5.2 Survival Distribution 9
3.5.3 Non-Parametric Methods 10
3.5.4 Cox-Regression Model
3.5.5 Partial Likelihood Function for Survival Times Without Tied Survival Times
3.5.6 Accelerated Failure Time Model 12
Chapter Four
Results
4.1 Introduction
4.2 Survival Data Summary
4.3 Log Rank Test 16
4.4 Cox Proportion Hazard 19
4.5 Accelerated Failure Time 21
4.6 Comparison 21
Chapter Five
Conclusion
5.1 Introduction
5.2 Limitations 25
5.3 Conclusion
5.4 Recommendations
Appendix 27
Bibliography 32
Chapter One
Introduction
1.1 Background Information
The field of statistics known as survival analysis examines the desired outcome until the occurrence of the event. As a result, it’s referred to be "time to event analysis," employed in a variety of fields including medicine, manufacturing, transportation, e-commerce, hu- man resources, and engineering. When a patient is diagnosed with a terminal condition such as cancer, medical science understands how long they will live. In the manufacturing sector, time to events such as when a car battery dies permanently is of interest. When it comes to predicting when an employee will quit a job and measuring employee retention and satisfaction, human resources can help. Given that the focus is on machine model- ing or components linked with electronics, engineering has participated in the research of survival analysis known as "failure time analysis." As a result, different advancements in the field of survival analysis have been incorporated into various fields. However, there are minor differences in the techniques utilized, such as duration analysis in economics.
Complications stemming from censored observations infiltrated statistical methodology developed primarily in the second part of the twentieth century. We will concentrate on frequentist methods in our application, despite the fact that Bayesian Survival approaches [14] have been substantially established and are growing in popularity for survival data. A number of textbooks have been developed on the same: Lawless [18], as well as Oakes [6], Fleming and Harrington [10], and Klein and Moeschberger [15] are only a few examples.
Kaplan and Meier [16] made a significant contribution to the non-parametric analytic method. They function well with similar class samples, but don’t assess any particular variables are linked to survival duration. But since survivorship durations are very seldom distributed normally, and redacted data results in lacking survivorship times, this loop- hole leads to regression method application, but it still falls short in survival data. The Cox PH model is frequently included in survival analytical data processes when variables are included because of its ease. The underlying hazard rate, according to the model, is a function of the uncorrelated variables and not of the hazard function’s shape. The original has been extended through modifications.
The semi-parametric method has no assumptions about how the event’s underlying risk evolves over period of time therefore the Cox PH model is more largely used compared to parametric methods for analyzing time-to-event data. Exponential, lognormal, Weibull, log-logistic and other hazard distributions are examples of hazard distributions. The relative hazard is calculated using both semi parametric and parametric approaches. Modeling of actual failure times is possible with some distributions. During the nth percentile occurrence of individuals is achieved, The result of an intrigue could be explicitly quantified via the accelerated failure models’ as a measure of connection. It is anticipated that the fundamental risk will match a Weibull. using time-to-event data.
1.2 Problem Statement
Although Cox PH together with AFT models have been widely utilized in survival time prediction, a dilemma limits the accuracy of these prediction methods. Limited collection sizes and censored input continue to be a barrier to training reliable and precise models Cox classification models. Despite all these, AFT models are least used in employee attri- tion. As a result, the goal of this research is to conduct survival analysis on the employee attrition data set to narrow down on the specific factors that will benefit the employer.
1.3 Objectives
Main Objective:
To compare AFT to Cox PH models in the employee attrition to determine “survival” of employees.
Specific Objectives:
1) To identify factors affecting employee attrition using survival analysis.
2) To compare survival probabilities obtained using AFT and Cox PH.
3) To obtain the model appropriate model for the data.
1.4 Significance of the Study
The study findings would reveal the risk variables or the most important covariates that have a substantial impact on employee attrition. A number of characteristics will be ex- amined in this study, including gender, overtime, business travel, and status of marriage, to mention a few. The research will assist in identifying the risk of employee attrition in the involvement of major circumstances. The findings will also provide a better under- standing of how to apply the concepts of standard measure of variability and AIC to the data set.
As a result, offering solutions that can predict staff turnover could be extremely useful to businesses. Furthermore, by combining Survival Analysis with the temporal dimension, it is possible to anticipate when an employee will leave.
1.5 Scope of the Study
While survival analysis has been well documented in various fields, the employee attrition analysis is quite limited. Using a fictional data set of 1470 employees with different attributes that will assist to focus the research within achievable parameters. This is a fictional data set with 9 attributes: 1) Business travel involved 2) Role of the employee in the organization 3) Levels of Work Involvement: Low, Medium, High, and Very High 4) Marital status of the employee 5. Distance between working place and home 6) Yes or No to overtime 7) Poor, Good, Better, Better Working Harmony 8) Gender of the employee 9) Level of satisfactory in the job
Click “DOWNLOAD NOW” below to get the complete Projects
FOR QUICK HELP CHAT WITH US NOW!
+(234) 0814 780 1594
Buyers has the right to create
dispute within seven (7) days of purchase for 100% refund request when
you experience issue with the file received.
Dispute can only be created when
you receive a corrupt file, a wrong file or irregularities in the table of
contents and content of the file you received.
ProjectShelve.com shall either
provide the appropriate file within 48hrs or
send refund excluding your bank transaction charges. Term and
Conditions are applied.
Buyers are expected to confirm
that the material you are paying for is available on our website
ProjectShelve.com and you have selected the right material, you have also gone
through the preliminary pages and it interests you before payment. DO NOT MAKE
BANK PAYMENT IF YOUR TOPIC IS NOT ON THE WEBSITE.
In case of payment for a
material not available on ProjectShelve.com, the management of
ProjectShelve.com has the right to keep your money until you send a topic that
is available on our website within 48 hours.
You cannot change topic after
receiving material of the topic you ordered and paid for.
Login To Comment