Lecture 1A: Statistical Estimators of Grade: Min 4025 Geostatistics
Lecture 1A: Statistical Estimators of Grade: Min 4025 Geostatistics
Lecture 1A: Statistical Estimators of Grade: Min 4025 Geostatistics
SCHOOL OF MINES
DEPARTMENT OF MINING ENGINEERING
M IN 4025 1
TOPIC 1
STATISTICAL ESTIMATORS OF THE GRADE
Objective: Review Basic Statistics
1
STATISTICAL ESTIMATORS OF THE GRADE
Figure 1:
2
STEPS IN APPLICATION OF STATS
1. Database
2. Data analysis
3. Frequency distributions
4. Estimation of mean and variance
5. Calculation of confidence interval for the mean
3
SHAPE OF ASSAY DISTRIBUTION
• Kurtosis = ((Xi-m)4)/ 2
Where;
= Variance of values
- Close to three for a normal
distribution
- Measure of ‘peakness’
4
COEFFICIENT OF VARIATION (COV)
10
5
Use log probability plot to test for lognormality
Resource Estimation & M in Geo 311 11
- Aplication of Statistics
11
12
12
6
Direct back transformation of log mean yields an estimate of the
median
13
13
14
7
SICHEL’S MEAN
log variance
Sichels' s Mean = e (log mean)
e 2
Where;
Log mean = Mean of transformed data (Log10(Xi)
Log variance = Variance of transformed data
15
TOP CUTS
16
8
Top Cutting means resetting all high samples to the top cut value
17
17
18
9
Assess grade of disintegration (Snowden, 2004)
19
19
20
20
10
CUMULATIVE DISTRIBUTIONS (CDF)
21
21
22
22
11
DECLUSTERING
23
24
24
12
25
25
26
26
13
Example data set:
27
% CaF2
Mean of class interval
Resource Estimation & M in Geo 311 28
- Aplication of Statistics
28
14
ESIMATION OF MEAN AND VARIANCE
• Mean = ( Xi)/n
• Consider sample support:
- Mean = ( ai*Xi)/ ai
- ai may be core length
• Mean of data within class width
- Mean = (( fi*Xi)/ fi
- fi = frequency of class interval
- Xi = mean of class interval
29
29
30
15
Figure 2:
31
31
32
32
16
Confidence Interval for the mean
33
33
• Sample mean,
x
1
x= i
n i =1
34
34
17
Sample Variance:
2
()
n
1
s =
2
xi − n x
n − 1 i =1
35
35
• Standard deviation:
s = s2
• Mean and variance of ore body:
=x
s2
V ( ) =
n
36
18
Resource Estimation & M in Geo 311 37
- Aplication of Statistics
37
38
38
19
• Let p = confidence limit of true mean such that the
probability that is smaller than p is p.
• Then 1-p is the confidence limit such that the probability that
> 1- is p
• The probability that falls between up and 1-p is (1-2p) and the
limits are called the central (1-2p) confidence limits for the mean.
• If n is greater than 25, the following approximation equations
can be used to calculate the central 68% and 95% confidence
limits for the mean value of the orebody (ie, p= 16% and p = 2.5%
respectively)
39
39
40
40
20
CONFIDENCE LIMITS FOR ‘n’ GREATER THAN 25:
s s
68% CI: x− ,x+
n n
s s
95% CI: x−2 ,x+2
n n
41
41
s
• Lower Limit: p = − t1− p
n
s
• Upper Limit: 1− p = + t1− p
n
42
42
21
Where;
43
43
EXAMPLE 1
• Eight samples have been taken from an orebody, whose
values are assumed to be normally distributed.
• Sample values:
i 1 2 3 4 5 6 7 8
Xi 1.2 2.0 1.6 1.7 2.5 1.9 1.5 2.1
44
44
22
SOLUTION
• Mean = 1.812
• Variance = 0.161
• Standard deviation = 0.402
• Given 1-p = 0.90, we calculate p = 0.05 and for n=8 we
derive from Table 2.2, t 0.95 = 1.895
• Hence Confidence Limits for the mean:
-Upper Limit = 1.812 + 1.895 (0.402/81/2 = 2.08
-Lower Limit = 1.812 - 1.895 (0.402/81/2 = 1.54
45
45
Exercise:
• DO Ex.2.9, Question 9 on Pg 27 of the book “Geostatistics
2000; Available on WhatsApp group/Moodle
• Students to work in pairs
46
46
23