This project was conducted as part of the Advanced Machine Learning course, a component of our Bachelor of Science studies in Business Informatics at Baden-Württemberg Cooperative State University. It focuses on utilizing Machine Learning to predict job change readiness among participants of Data Science training programs. Our goal was to develop models that can accurately identify individuals likely to switch jobs, potentially aiding recruitment processes.
Amaan Ansari, Robin Bischoff, Nina Resch
The primary objective was to analyze a dataset of training program participants and predict their likelihood of job change. We employed various classification models, optimizing for prediction accuracy.
The dataset, sourced from Kaggle, includes various features such as City Development Index, Gender, Relevant Experience, Education Level, and more. The target variable indicates whether an individual is seeking a new job (1) or not (0).
Key steps in data preparation included:
- Handling missing values and categorical variables Addressing class imbalance
- Feature engineering for optimal model performance
We employed both basic and complex models:
- Basic Models: Logistic Regression and Decision Tree
- Complex Models: Boosting, Random Forest, and Support Vector Machine
Models were evaluated based on their F1-Score, with a focus on maximizing the prediction accuracy for individuals seeking new job opportunities.
Our analysis revealed that complex models did not significantly outperform the simple Decision Tree model. The City Development Index and Company Size emerged as the most influential variables in predicting job change readiness.
The project highlights the importance of data preparation and model selection in predictive analytics. Despite the modest overall performance, the models provide valuable insights into the factors influencing job change decisions among training participants.
Grade: 1.1
For those interested in a deeper understanding of our methodologies, findings, and analyses, the full project report is available upon request. Please feel free to reach out to me for more detailed information and insights from our work.