Student blog: Federica Lucia Vinella
In the next instalment of our student blog series we hear from Federica Lucia Vinella who is also doing a placement at Wallscope while studying for her MSc in Computing at Edinburgh Napier University.
I am currently developing a dissertation on data mining in medical text. The topic was inspired and matured as part of the summer work placement at Wallscope and relates to a project the company is currently working on in collaboration with the Information Services Division of NHS Scotland.
I came across Wallscope after researching for placements within CodeBase. The proposed project to detect prescription drug names from medical text meant that I would have to utilise the data wrangling knowledge gained during my course whilst learning new methodologies and industry-based insights to address the task.
During the placement I have been working on generating two deep learning algorithms for drug Named Entity Recognition with the implementation of SpaCy and Recurrent Neural Networks. One of the greatest challenges of data mining for medical text is the acquisition and generation of accurate annotated datasets, followed by the prediction of drug entities that can be classified by their generic, proprietary and group names. The complex nature of the annotation with the specifications of the medical domain made the project particularly interesting to me.
There is a vast amount of unstructured data regarding medical records that can be used to optimise information extraction in relation to drug-drug interaction, drug administration and statistical analysis. With the use of machine learning algorithms it’s possible to make semi-accurate predictions of where the drugs are used as well as the ability to identify drug names in contextual settings. By providing an automated tool for text processing for the medical staff and health institutions to use, it could be possible to improve drug surveillance and deepen the understanding of drug use across Scotland.
After my MSc I am hoping to start a career as a junior data analyst and expand my knowledge in the field of Python programming, Artificial Intelligence, machine learning and text mining. I think that working for a company as part of a placement is by far the best investment for the future as it prepares for real-life challenges and newer technologies that are not always investigated in an academic environment. Another very important aspect is being able to communicate and learn from people who have furthered their experience and can massively help tackling problems and advise on insightful solutions. I have learnt a great deal from the placement and I would highly recommend it to every student.