Earlier this year, we welcomed school students to our Lincoln office to show them the game changing potential that Industry 4.0 technologies are having for businesses. We set the students a challenge of using machine learning to make sense of software engineering data collected from MASS software products.
We provided them with a data set containing software metrics relating to size, coupling, and complexity for each class in the codebase, as well as a measure derived from historic observation reports indicating whether each class is error-prone or not. The students were shown how to write Python machine learning code using Jupyter Notebooks hosted in Google’s Colab Laborator. Once briefed, the students worked as a team to train machine learning algorithms to predict which classes are likely to be error-prone based on the metrics provided.
The students decided to experiment with a selection of algorithms including Linear Regression, Decision Trees, and Neural Networks. It was great to see how they keenly collaborated through pair programming and diligently researched more about how the algorithms work using online resources. By the end of the week, they had managed to optimise the algorithms and improve output reporting. We were delighted with the results they achieved and impressed with the insightful questions they asked along the way.
Logan Lynch one of the school students who attended the work experience at our offices said of his experience, “I am grateful to have had the opportunity to work on this project. I enjoyed the week at MASS and the opportunity to work with and learn from experienced software developers. I had never previously experienced machine learning and cannot wait to know more about it in the future.”
He concluded by saying, “This experience also showed me what a real-life working environment looks like and will prepare me for future opportunities in years to come.”
If you’re interested to learn more about using Python for machine learning, the links provided below provide a great starting point, and Google Colab is a highly accessible and easy to use programming environment that can be used with nothing more than a standard web browser.