c173c273_lec9_w11[1]

c173c273_lec9_w11[1] - University of California, Los...

Info iconThis preview shows pages 1–5. Sign up to view the full content.

View Full Document Right Arrow Icon
University of California, Los Angeles Department of Statistics Statistics C173/C273 Instructor: Nicolas Christou Spatial statistics - prediction Spatial prediction: a. One of our goals in geostatistics is to predict values at unsampled loacations. This will create a “raster map” of the area under consideration. b. Idea: The value at an unsampled location is related to the values at sampled locations. If not, then the predicted value can be simply the average of the sampled locations. c. There is some relationship between the point being estimated and the samples around it based on its location and the distance between the points. d. Different methods for spatial prediction. All methods assume that nearby samples are more important (they have more weight) in predicting the unknown value. Some simple methods of spatial prediction: 1. Method of polygons. 2. Triangulation. 3. Weighted average based on triangulation. 4. Inverse distance method. A. Method of polygons: Very simple. The point being estimated is equal to the nearest observed sampled data point. 1
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
B. Triangulation Form a triangle around the point being estimated with three observed data points. For these three data points we know their x,y coordinates and the value z of the variable of interest as follows: x y z x 1 y 1 z 1 x 2 y 2 z 2 x 3 y 3 z 3 We use these points to find the equation of a plane by solving the system of equations: ax 1 + by 1 + c = z 1 ax 2 + by 2 + c = z 2 ax 3 + by 3 + c = z 3 Suppose now we want to find the value of z 0 at location given by the coordinates ( x 0 ,y 0 ) of a point within the triangle. Example: Estimate the value at (65, 137) if the surrounding three points are: z 1 = 696 at (63, 140), z 2 = 227 at (64, 129), and z 3 = 606 at (71, 140). 2
Background image of page 2
C. Weighted average based on triangulation: Suppose we want to estimate the value at point O based on the values z A ,z B ,z C from the figure below: A weighted average is used as follows: ˆ z O = z A × Area( OBC ) Area( ABC ) + z B × Area( OCA ) Area( ABC ) + z C × Area( OAB ) Area( ABC ) This method gives more weight to the data point closer to the point being estimated. 3
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
D. Inverse distance method: Suppose we want to estimate data point from the observed points as shown in the figure below: Each observed data point has an impact on the point being predicted with the clos- est points having larger impact. We can consider as a predictor the following linear combination (weighted average): ˆ z ( s 0 ) = w 1 z 1 + w 2 z 2 + ··· + w n z n = n X i =1 w i z i The weights w i ,i = 1 , ··· ,n are inversely related to the distance of point i to the point being estimated. For example,
Background image of page 4
Image of page 5
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 02/11/2012 for the course STATS c173/c273 taught by Professor Nicolaschristou during the Spring '11 term at UCLA.

Page1 / 13

c173c273_lec9_w11[1] - University of California, Los...

This preview shows document pages 1 - 5. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online