Scribd
Upload
166 views3 pages

Assessment 1 - UEL-CN-7000

Uploaded by

SajjadJamil
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
166 views3 pages

Assessment 1 - UEL-CN-7000

Uploaded by

SajjadJamil
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Equitable and Transparent Salary

Prediction: Leveraging Machine


Learning for Demographic and
Professional Insights
Abstract
This research proposal aims to develop a more equitable and transparent salary prediction system
by leveraging machine learning techniques, particularly focusing on demographic and
professional factors. The need for fairer systems has grown, as existing predictive models have
been shown to exhibit biases based on gender, race, and age, leading to inequitable salary
forecasts (Blau and Kahn 2017). This research aims to use advanced machine learning
algorithms to create a more inclusive and accurate model for salary prediction. Through a
combination of demographic and professional insights, this research seeks to establish a robust,
transparent, and non-discriminatory prediction model that addresses current limitations
(Woodworth, Gunasekar et al. 2017) (Kasy & Abebe, 2019).

Table of Contents
1. Introduction
2. Preliminary Literature Review
3. Experimental Design and Methods
4. Proposed Timetable
5. References

Introduction
Salary prediction using machine learning has gained significant traction in recent years, but
challenges persist, particularly regarding fairness and transparency (Hartmann and Slapničar
2012). While machine learning models have been employed to predict salaries based on
professional and demographic factors, these models often fail to account for inherent biases in
demographic variables such as gender, age, and race. This bias can lead to inaccurate and
discriminatory outcomes, which affects individuals in marginalized groups (Blau and Kahn
2017).

The purpose of this research is to design a salary prediction model that ensures fairness by
incorporating demographic and professional data, while also ensuring the transparency of the
model’s decisions. Through this, the research will contribute to improving the fairness and
accuracy of salary prediction models. The research will be guided by the theoretical framework
of fairness in machine learning, as proposed by (Woodworth, Gunasekar et al. 2017)Kasy and
Abebe (2019), which advocates for developing models that are both accurate and unbiased in
their predictions.

This research is significant as it addresses the increasing concerns about algorithmic bias in AI
systems used in hiring and salary determination. As organizations increasingly adopt AI for
human resources decisions, ensuring that these systems are equitable is essential (Blau and Kahn
2017).

Preliminary Literature Review


Existing literature on salary prediction models reveals a growing body of work focused on
machine learning applications in human resources. Traditional models used for salary predictions
often overlook the impact of demographic variables, leading to biased predictions (Hartmann and
Slapničar 2012). For instance, gender and racial biases are prevalent in machine learning
systems, which are often perpetuated by biased historical data used to train these models (Blau
and Kahn 2017).

While some studies focus on improving model accuracy by incorporating more data, others
emphasize the importance of transparency and fairness (Woodworth, Gunasekar et al. 2017).
However, many studies neglect the ethical implications of using demographic data, failing to
address how these models can reinforce existing social biases (Blau and Kahn 2017).

The proposed research aims to bridge the gap in existing knowledge by combining both technical
and ethical approaches to create an equitable salary prediction model. By leveraging machine
learning algorithms and applying fairness constraints, this research will address the biases
present in current systems and offer a more inclusive solution (Woodworth, Gunasekar et al.
2017).

Experimental Design and Methods


This research will adopt an experimental design focused on leveraging machine learning
algorithms to predict salaries while accounting for demographic and professional data. The main
methodology involves using Random Forest, a machine learning technique known for its ability
to handle complex datasets with multiple demographic features (Patil, Rane et al. 2024). Data
will be collected from publicly available datasets containing salary information, along with
demographic details such as age, gender, and education.

To ensure fairness, the model will be trained with fairness constraints as suggested by
(Woodworth, Gunasekar et al. 2017), which aim to reduce bias in predictions by adjusting for
demographic discrepancies. Data analysis will involve evaluating the performance of the model
using standard metrics such as accuracy, fairness, and transparency.

Challenges may arise in terms of obtaining high-quality, unbiased data, as well as dealing with
issues such as overfitting in machine learning models (Patil, Rane et al. 2024). These challenges
will be addressed by employing cross-validation techniques and adjusting for bias in the training
dataset.

Proposed Timetable

Activity Duration Deliverable

Data Collection 1 month Compiled dataset of salary and demographic data

Data Preprocessing 1 month Cleaned and processed dataset

Model Development 2 months Trained machine learning model

Fairness Adjustment 1 month Model adjusted for fairness constraints

Results Analysis 1 month Performance evaluation report

Final Report Writing 1 month Completed research paper

References
Blau, F. D. and L. M. Kahn (2017). "The gender wage gap: Extent, trends, and explanations." Journal of
economic literature 55(3): 789-865.

Hartmann, F. and S. Slapničar (2012). "Pay fairness and intrinsic motivation: the role of pay
transparency." The International journal of human resource management 23(20): 4283-4300.

Patil, D., et al. (2024). "Machine learning and deep learning: Methods, techniques, applications,
challenges, and future research opportunities." Trustworthy Artificial Intelligence in Industry and
Society: 28-81.

Woodworth, B., et al. (2017). Learning non-discriminatory predictors. Conference on Learning Theory,
PMLR.

You might also like