Data Science Intern at Co-op Insurance
Additional Description of Tasks
Through the use of R and Python, use the company’s membership data in order to determine the best insurance price for their members, taking into account how likely a member is to default/crash
Explore different data science techniques to see which methods work best with an emphasis on customer interpretability
Improve upon the pre-existing methods that have been used in the past such as using non-parametric boosting algorithms such as XGBoost, random forests, and other machine learning techniques
Implement several hyper-parameter optimization techniques such as Bayesian Optimization (Gaussian Process - Upper Confidence Bound acquisition function), Random Search, and Grid Search
Produce clean, re-usable, and production-level code that can be used in the future
Produce a final dissertation that includes previous relevant works, methodologies, findings, results, and conclusions
Explore different data science techniques to see which methods work best with an emphasis on customer interpretability
Improve upon the pre-existing methods that have been used in the past such as using non-parametric boosting algorithms such as XGBoost, random forests, and other machine learning techniques
Implement several hyper-parameter optimization techniques such as Bayesian Optimization (Gaussian Process - Upper Confidence Bound acquisition function), Random Search, and Grid Search
Produce clean, re-usable, and production-level code that can be used in the future
Produce a final dissertation that includes previous relevant works, methodologies, findings, results, and conclusions
- Location: Manchester, United Kingdom
- Start Date: June 2018
- End Date: September 2019
- List of Relevant Technologies Used: R, XGBoost, rBayesianOptimization