DREaM event 4 speaker insight: Kevin Swingler
April 18, 2012 Leave a comment
The second in our series of preview interviews ahead of DREaM event 4 is with Kevin Swingler from Stirling University. In this interview, he introduces us to his topic of data mining and reflects how this might be applied to LIS research.
Which research method will you be discussing with the workshop participants in your presentation?
I will be discussing data mining techniques. These are methods for using data to ‘teach’ a computer to perform a task. Data mining is less concerned with understanding the data or the process that produced it than most techniques. In this sense it is task oriented – we use the data to predict future events or classify situations as being similar to those seen in the past.
How have you used this in your own research?
My own research includes devising new data mining techniques and using existing ones – mostly for commercial applications such as predicting consumer behaviour, spotting fraud in banking or insurance, and predicting medical outcomes.
How do you think this might be useful as a method in LIS research?
An example of where data mining might be useful in LIS is automatic sentiment classification in social media. This is the process of training a computer to tell whether the attitude in a social media message is positive or negative. It can also be used to find posts on certain topics where simple keyword lists are not enough.
Where can people will find more information?
I have a series of lecture slides on data mining from my course at Stirling University. You can see them here.
A good book is Data Mining: Practical Machine Learning Tools and Techniques by I.H. Witten and E. Frank
Kevin Swingler will be presenting a session introducing data mining at the fourth DREaM workshop at the Edinburgh Napier University on Wednesday 25th April. For full details about the workshop, please see the workshop programme.