MineMatch Reports

MineMatch records the kind of match recognised between each matching pair of target and deposit model attributes.

This assists the reader of the report to understand MineMatch's reasoning, and it allows MineMatch to adjust its similarity scores according to the strength or weakness of the match.

The following paragraphs explain the different kinds of matches recognised by MineMatch, with examples of each from targets identified in the Yukon.

Exact, Exact and Possible, Exact

The "exact, exact" (bright green) match arises when the two compared attribute-value pairs are exactly the same, as in the case of "ElementEnhanced - Sb" compared to "ElementEnhanced - Sb" in the screenshot in the right (click to enlarge).

A "possible, exact" (light green) matches arise when the attributes only possibly match each other, and their values match exactly (as do Au and Au).

This situation occurs, in respect of the "possible", when the compared attributes exist in a hierarchical "kind-of" relationship to one another, and the subordinate attribute (ie: of the target) is higher in the hierarchy than the dominant attribute (of the model). In the example above, an "ElementEnhancedToOre" is a kind of an "ElementEnhanced", but we know only that an enhanced level of the element (higher in the hierarchy) exists in the anomaly cluster, not that it is definitely "enhanced to ore levels", as demanded by the model.

Exact, AKO

The Exact, AKO (bright green) match arises when the compared attributes are exactly the same, and the value of the subordinate participant in the comparison is a kind of the value to which it is being compared.

When a target is being compared to a mineral deposit model, the target is the subordinate participant in the comparison.

In the example above (click to enlarge), a dolostone is a kind of carbonate sediment, and an arc system is a part of an accretionary prism. Therefore, in the current MineMatch taxonomy of tectonic settings (adopted from that proposed by Prof. Kent Condie), an arc system is reasoned with as a kind of accretionary prism. (Work is currently in progress in Canada and the USA on formalising a taxonomy of tectonic settings, and once this is complete, the taxonomy will be incorporated into MineMatch.)

Possible, Possible

"Possible, Possible" (light green) matches arise when the compared attributes only possibly match each other, and their values may also only possibly match.

This situation occurs, in respect of the first "possible", when attributes exist in a hierarchical "kind-of" relationship to one another, and the subordinate attribute (ie: of the target) is higher in the hierarchy than the dominant attribute (of the model). In the example above, a "locating form" is a kind of a "form", but we know only that a form (higher in the hierarchy) exists close the anomaly cluster, not that it is definitely a "locating form", as demanded by the model.

Similar considerations apply to the attribute values. "Faults" constitutes a kind of faulting, and is therefore lower in the taxonomy of form than "faulting". But it is the attribute of the dominant object in the comparison, and therefore falls into the "possible" category of match in this context.

Further explanation of the issues which have to be addressed in providing these kinds of explanation can be found in the following reference:

Qualitative Probabilistic Matching with Hierarchical Descriptions
by Clinton Smyth and David Poole,  KR-2004, June 2004.
 (A summary of the paper is also available as Powerpoint slides.)