Currently I am busy with my MSc Computer Science while co-supervised by the Archaeology Department. I am in the Computational Intelligence Research Group, so my research is focused on Computational Intelligence.
What relevance does this have to Computational Archaeology? Well, archaeologists deal with large amounts of documentation, often so large that it is difficult to accurately interpret a scenario by taking all the information into account. This is where computers come in. Computers can handle large portions of data at a time in some predefined way. Data Mining using Computational Intelligence methods allows for the computer to process the data without any prior “knowledge” on the dataset.
My research topic is “The clustering of temporal archaeological data”. This means that I will be given two temporal archaeological datasets on which I will apply a set of Computational Intelligence algorithms in order to:
- test the algorithm’s performance with temporal datasets and
- find out interesting relationships among the patterns in the archaeological dataset.
I will be testing 3 algorithms. Each algorithm has a set of variations of which some will be tested. So far 6 algorithms of one type have been implemented and experiments are being done with these algorithms.
The obvious advantage of using computational intelligence to mine archaeological data is the load taken off the archaeologist. The archaeologist can focus on his/her archaeological tasks more, rather than spend time with administrative issues. It also removes human bias from the initial process of interpreting the data. The archaeologist will still deal with the data, but only after it has been processed to make the interpretation easier. With my research, similar data patterns will be identified as separate groups. These groups can then be analysed by the archaeologist in order to determine why the algorithm found the patterns in these groupings to be similar, possibly giving the archaeologist a different perspective on the data.
Visualisation of the data may be helpful when there are few attributes to consider, however, once the dimensionality increases, it is a lot more difficult to visualise the data in an understandable way. The computational intelligence algorithms that I am busy with can give the archaeologist sets of data found to be similar, allowing the archaeologist to see relationships which he/she may have missed when interpreting the original dataset.
I started my research at the beginning of this year and hope to finish around the middle of next year.