Transiogram

From Wikipedia, the free encyclopedia

Transiogram is the accompanying spatial correlation measure of Markov chain random fields and an important part of Markov chain geostatistics. It is defined as a transition probability function over the distance lag. Simply, a transiogram refers to a transition probability diagram. Transiograms include auto-transiograms and cross-transiograms. The former represent the spatial auto-correlation of a single category, and the latter represent the spatial interclass relationships among different categories. Experimental transiograms can be directly estimated from sparse sample data. Transiogram models, which provide transition probabilities at any lags for Markov chain modeling, can be further acquired through model fitting of experimental transiograms.