slice_instructions3

# slice_instructions3 - To check on the effectiveness of the...

This preview shows pages 1–4. Sign up to view the full content.

To check on the effectiveness of the conditional estimation procedures, it is convenient to slice through the modes of the two rightmost of the three normal density components. To get the slice as close to the rightmost modes as possible, put the pointer over the middle of the right hand red cluster of the previously displayed stickpin map. Now click a mouse button to draw the slice. (Your slice should now appear over a value close to x = 3.3.) The screen will look as shown above. To refine the curve parsing process still further, select the Return box to go back to the main menu. Then choose the CONTRST option. From the Contrast submenu select the Lambda box. At the Enter Lambda: prompt, type in 1.1, and press the Enter key. The following screen will appear: FIGURE 10. CONTRAST REDUCED CURVE Next choose the option Separate to decompose the distribution into two components. A message at the bottom of the screen will prompt you to select two points (by using the mouse) in succession. To isolate the left component of the density, first move the pointer to the far left side of the X axis (approximately at x = -5) and press a mouse button. Next move the pointer to the trough between the two highest modes (near x = 0) and press the mouse button. The portion of the density that lies between the two chosen points will isolated mathematically from the remainder of the density and will be displayed using the original horizontal scale. The estimated mixing parameter (in per cent) will be shown at the bottom of the screen, 9

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

View Full Document
and should be close to 50%. (The two rightmost components were chosen to be of equal size 30% and 30%, respectively, of the entire distribution. Since they are of equal size, each slice component comprises 50% of the total slice.) Place the pointer on the Get XY box and click a mouse button. Use the mouse to move the pointer to the very top of the estimated density curve; then press the mouse button again. The (X,Y) coordinates of the point you just selected will appear in a message at the bottom of the screen. The X value, the estimated mode of the separated component, will be about -1.3, and the Y value, the estimated height of the density at the mode, will be about .36. (The mean of the leftmost density component is -1.2.) A picture of how the screen will look is shown below: FIGURE 11. SEPARATED CURVE Select Return to return to the main menu and then select the R(x) box. When the submenu appears on the left side of the screen, place the pointer over the Normal box and click the mouse. A checkmark appears in the Normal box to show that the Normal R(x) curve has been selected, but the curve will not be drawn until the Redraw box is selected. To specify the region within which the R(x) curve will be computed, select the Set Range box and use the same box specification procedure that was use to specify the lengths of displayed regression curves. After you have done this, and before you select Redraw , the screen will look like this: 10
FIGURE 12. RANGE SPECIFICATION INTERVAL FIGURE 12. ESTIMATED R(x) CURVE AND ESTIMATED THRESHOLD PARAMETER 11

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

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 02/04/2011 for the course PB HLTH 140 taught by Professor Tarter during the Fall '10 term at University of California, Berkeley.

### Page1 / 9

slice_instructions3 - To check on the effectiveness of the...

This preview shows document pages 1 - 4. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online