Objective
The objective of this study is to investigate the possibility of mitigating discriminatory biases in decision-making using a decision tree, while also providing explainability to the algorithm’s classification.
Specifically, the experiment will be conducted in the context of personnel selection and the gender variable, although the aim is to generalize to other areas as well.
In addition to seeking the reduction of biases, it is also intended to reflect the real-world situation given the accuracy of the model In this way, the selection personnel can indicate the desired distribution and the importance of that consideration compared to reality. For example, the HR staff may wish to have a 50% distribution of women in the workforce, but they also want to choose the most suitable candidate, so they can indicate that gender is 80% important, allowing room for real classification.
Finally, the algorithm should return a list of candidates classified as “to interview” or “not to interview.” Among those indicated for an interview, they must comply with the indicated distribution.
BACHELOR’S THESIS BY:
JIA JIA YE
Academic Experience
Double Degree in Computer Science and Engineering, and Business Administration, Universidad Carlos III de Madrid
(september 2017 – june 2023)
Bachillerato de Ciencias Sociales (2015 – 2017)
Work Experience
Data Analyst – Grupo MasMóvil (julio 2023 – current)
Frontend Developer Intern – Grupo MasMóvil (december 2022 – june 2023)
Machine Learning Researcher – Universidad Carlos III de Madrid in collaboration with Grupo MasMóvil (september 2022 – may 2023)
RPA Department – NTT Data (january 2022 – august 2022)
Technical skills
Programming languages: Python, Java, C/C++, SQL, HTML/CSS/JS
Relevant libraries: Pandas, NumPy y Scikit-Learn
Cloud Platforms: Google Cloud.
Other technologies and frameworks: Git, Hugo, Angular