PARM--an efficient algorithm to mine association rules from spatial data

Area of Science:

• Data Mining
• Geospatial Analysis
• Remote Sensing

Background:

• Association rule mining is valuable for pattern discovery.
• Spatial data, especially remote sensed imagery (RSI), presents unique challenges due to large sizes.
• Existing algorithms struggle with the scale of spatial datasets.

Purpose of the Study:

• To propose an efficient algorithm for association rule mining on large spatial datasets.
• To leverage the Peano Count Tree (P-tree) structure for data compression and efficient mining.
• To improve the speed and effectiveness of extracting rules from RSI data.

Main Methods:

• Developed the P-tree based Association Rule Mining (PARM) algorithm.
• Utilized P-trees for lossless compression and fast support calculation.
• Incorporated pruning techniques to enhance mining efficiency.

Main Results:

• PARM demonstrated superior performance compared to FP-growth and Apriori algorithms.
• The P-tree structure enabled efficient handling of large RSI datasets.
• Fast support calculation and pruning significantly improved the mining process.

Conclusions:

• The PARM algorithm offers an efficient solution for association rule mining in large spatial datasets.
• P-trees are effective for compressing and mining RSI data.
• PARM is a promising approach for applications in precision agriculture and resource discovery.