Abstract:

Application of the Weighted Logistic Regression model in prediction of volcanic rock type Copper deposits in the Middle
part of Ning-Wu Basin is studied. First, the geological setting of ore-forming processes is analyzed. Three kinds of factors including
geological body, structure and wall rock alteration are extracted based on the spatial distribution of copper deposits from the geologic
map. Then, the spatial relationships between Copper mineral occurrence and each evidence factor are analyzed. It is suggested that
Niangniangshan and Gushan volcanic edifice play an important role in spatial distributions of volcanic rock-hosted Copper deposits.
The ten evidence raster layers including Longwangshan Formation, Gushan Formation, trachyte porphyry of Gushan volcanic edifice,
monzonite porphyry of Niangniangshan volcanic edifice, buffers of the structure lines with NE, NW and EW trending, and the alteration
areas of chalcopyrite, silicide and Limonite are selected. Finally, metallogenic probabilities are calculated using the Weighted Logistic
Regression model. Four ore-forming prospects, including P1, P2, P3 and P4, are indicated based on the geological conditions of

metallogenesis and model results. Among these prospecting areas, P1, P2 and P3, which are controlled by Niangniangshan and Gushan
volcanic edifice, are spread in the northeast direction. P4 extends in the west-east direction and is controlled by Longwangshan volcanic
edifice. The copper ore bodies are already found in these prospecting areas, suggesting that the results should be generally reliable.

Key words: weighted logistic regression model, Ning-Wu Basin, weight of evidence model, mineral prediction