Optimization and visual controls for regionalization
In accordance with certain embodiments of the present disclosure, a regionalization method is disclosed. The method includes inputting a data set into a computer. The method further includes utilizing the computer to perform contiguity-constrained hierarchical clustering on the data set to generate two regions and performing a fine-tuning procedure on the two regions with the computer to iteratively modify the boundaries between the two regions.
1. A regionalization method comprising:
inputting a data set into a computer;
utilizing the computer to perform contiguity-constrained hierarchical clustering on the data set to generate two regions while minimizing information loss associated with the regions, the information loss being defined in terms of the sum of squared distances; and
performing a fine-tuning optimization procedure on the two regions with the computer to iteratively modify the boundaries between the two regions while minimizing information loss by moving at least one data point from one region to the other while maintaining the contiguity of each region to enhance partition quality.
2. The method of claim 1 , further comprising:
performing the method iteratively to generate a hierarchy of regions.
3. The method of claim 1 , wherein the computer is configured to maintain spatial contiguity between regions.
4. The method of claim 1 , wherein the method is utilized for climate zoning, public health mapping, political redistricting, or combinations thereof.
5. The method of claim 1 , further comprising meeting multiple constraints during the fine-tuning procedure, the constraints comprising minimum region size, geographic contiguity, region shape, or combinations thereof.
6. The method of claim 1 , wherein the contiguity-constrained hierarchical clustering on the data set minimizes the information loss associated with each region.
7. The method of claim 1 , further comprising maintaining geographic contiguity by locating articulation points and allowing multiple-object moves.
8. The method of claim 1 , wherein the fine-tuning procedure modifies the boundaries between the two regions by moving multiple data points from one region to the other while maintaining the contiguity of each region.
9. A system for regionalization comprising:
a computer configured to receive a data set, the computer further configured to
a) perform contiguity-constrained hierarchical clustering on the data set to generate two regions while minimizing information loss associated with the regions, the information loss being defined in terms of the sum of squared distances,
b) perform a fine-tuning procedure on the two regions to iteratively modify the boundaries between the two regions while minimizing information loss by moving at least one data point from one region to the other while maintaining the contiguity of each region, and
c) iteratively perform the preceding two steps to generate a hierarchy of regions to optimize the objective function under constraints.
10. The system of claim 9 , wherein the computer is configured to perform contiguity constrained hierarchical clustering on the two regions to generate new regions and perform a fine-tuning procedure on the new regions to iteratively modify the boundaries between the new regions.
11. The system of claim 9 , wherein the computer is configured to maintain spatial contiguity between regions.
12. The system of claim 9 , wherein the computer is utilized for climate zoning, public health mapping, political redistricting or combinations thereof.
13. The system of claim 9 , further comprising an optimization procedure to find a sequence of moves to maximize the objective function and maintain contiguity and constraints.
14. The system of claim 9 , further comprising a visual interface to configure the parameters and examine the resulted regions.
15. The system of claim 9 , wherein the contiguity-constrained hierarchical Clustering on the data set minimizes the information loss associated with each region.
16. The system of claim 9 , wherein the fine-tuning procedure modifies the boundaries between the two regions by moving data points from one region to the other while maintaining the contiguity of each region.
17. The system of claim 9 , further comprising considering different Objective functions of a user's choice.