Thursday, 12 November 2009
Given weather data of various locations, grouping similar locations based on this data has always been an important part of meteorology. This grouping is most commonly known as clustering and is performed by assigning similar locations into subsets called clusters. Both hierarchical and partitional clustering techniques are used to cluster the data. A difference between the maximum temperatures at each location is the distance measure used for the clustering. The clustering is performed for every year and the result is a snapshot of what the climate was like for any particular year. Clustering historical data is important because we can observe how these clusters have changed over the years. From this, we can observe patterns than can hopefully be used to better predict climate change in the future.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner