1. Incorporating spatial association into statistical classifiers: local pattern-based prior tuning
- Author
-
Jinfeng Wang, Yong Ge, Qian Chen, Hexiang Bai, M. Peter Atkinson, and Feng Cao
- Subjects
Computer science ,business.industry ,Association (object-oriented programming) ,05 social sciences ,Geography, Planning and Development ,0211 other engineering and technologies ,0507 social and economic geography ,Pattern recognition ,02 engineering and technology ,Library and Information Sciences ,Class (biology) ,Local pattern ,ComputingMethodologies_PATTERNRECOGNITION ,Spatial ecology ,Common spatial pattern ,Classification methods ,Artificial intelligence ,business ,050703 geography ,Spatial analysis ,021101 geological & geomatics engineering ,Information Systems ,Statistical classifier - Abstract
This paper proposes a new classification method for spatial data by adjusting prior class probabilities according to local spatial patterns. First, the proposed method uses a classical statistical classifier to model training data. Second, the prior class probabilities are estimated according to the local spatial pattern and the classifier for each unseen object is adapted using the estimated prior probability. Finally, each unseen object is classified using its adapted classifier. Because the new method can be coupled with both generative and discriminant statistical classifiers, it performs generally more accurately than other methods for a variety of different spatial datasets. Experimental results show that this method has a lower prediction error than statistical classifiers that take no spatial information into account. Moreover, in the experiments, the new method also outperforms spatial auto-logistic regression and Markov random field-based methods when an appropriate estimate of local prior class distribution is used.
- Published
- 2020
- Full Text
- View/download PDF