Hi Abani, You have forgotten the last, and arguably the most important verification of your block estimates (although it is admittedly not a statistcal one): physically examine the block estimates in conjunction with the composite grades via plots, or the PC screen. Unfortunately, a pure statistical approach (histograms for example) to verification is non-spatial (even at the single domain level), and tells you very little about how well the estimates have honoured the spatial characteristics of your composites. I would urge caution in using the range of a variogram and average weighted distances as the sole basis for classifying resources. There is considerably more to resource classification than the inputs used to derive the estimates. See http://www.jorc.org/main.php for more information of classifying resource estimates. Regards, Colin
________________________________ From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Abani R Samal Sent: Saturday, 20 January 2007 2:19 AM To: [email protected] Subject: AI-GEOSTATS: Statistical analysis to verify BLOCK MODEL I have few questions about statistical analysis to verify how close is my estimates (BLOCK) are as compared to the input data i.e composited drill hole (DH) data. Please comment and suggest on the following questions/ assumptions. 1: My first approach: Make Histograms of both DH and BLOCK data and compare their statistics. My hypotheses are: Their means will be the same (or similar), so are their spread (variance and standard deviation). And at 90 - 95% Confidence Interval (CI) these parameters should be found to be same. Also I am expecting a close match of these parameters (irrespective of their distibution: normal/ log-normal) at different intervals of grades: for an examples, I expect these match should be valid at 0.5 - 1% Ni, 1-1.5% Ni, 2-3% Ni etc. Are these correct? If not what is the next best way to test thaese? 2: Secondly, if we do a Cumulative Distribution Function (CDF) plot, should the two match? If so, how close? And how and where we say "yes they are close enough" (?) . ......... should we try to match the Quantiles at certain CI? 3: Further is it justified to say that if narrower CI of block data (BLOCK) as compared to that of the sample (DH) data(for the whole data set OR, any particular interval as described above), means better the estimation? 4: In a resource classification scheme (Measured Indicated Inferred: MII), I was told to use 95% of the range of the variogram (D95) as the criteria for classifing the resource as Indicated resounce i.e, if D95/2 < Dist (Weighted average distance attached to each block estimated)<= D95, then the resource is Indicated AND if D95/2 >= Dist (Weighted average distance attached to each block estimated, then the resource is Meadured So the MII classifaction is done now. Now the question is: Is the D95 criteria provide the best MII classification? How do I test that? I guess it is a sensitivity analysis on the Block estimates. Please respond. Thanks in advance to all, who respond to my messages. Abani R Samal ************************************************************ ABANI RANJAN SAMAL 11183 West 17th Avenue, APt 201 Lakewood, CO 80215 ________________________________ 8:00? 8:25? 8:40? Find a flick <http://tools.search.yahoo.com/shortcuts/?fr=oni_on_mail&#news> in no time with theYahoo! Search movie showtime shortcut. <http://tools.search.yahoo.com/shortcuts/?fr=oni_on_mail&#news> This message and any attached files may contain information that is confidential and/or subject of legal privilege intended only for use by the intended recipient. If you are not the intended recipient or the person responsible for delivering the message to the intended recipient, be advised that you have received this message in error and that any dissemination, copying or use of this message or attachment is strictly forbidden, as is the disclosure of the information therein. If you have received this message in error please notify the sender immediately and delete the message.
