I currently need to do risk analysis data mining on a dataset. This dataset has around 120 attributes.
Although I can use common sense, is there any systematic methodology to do data reduction that it can guide us to choose which attributes are likely useful to feed into our algorithm?
What you're describing is feature or attribute selection. Weka does this from the "Select attributes" tab. You can find articles and videos about the topic online. I have found videos from the University of Waikato helpful. Here is one on attribute selection.
You may also want to learn about Principal Component Analysis.
