Euradwaste '08 - EU Bookshop - Europa
Euradwaste '08 - EU Bookshop - Europa Euradwaste '08 - EU Bookshop - Europa
2.2.3 Graphical methods Let us call X=(X1,X2,…,XK) the vector of input parameters and Y to a given scalar output variable. For a given input Xi, the scatter-plot is the projection of the sample points (X,Y) on the (Xi,Y) plane. This representation allows the examination of the dependence between Y and Xi. Scatterplots are very helpful to identify linear relations, monotonic relations and the existence of thresholds among other potential trends. A wise use of transformations (e.g. logarithmic, ranks, etc.) may also provide a lot of information about input/output relations. They may be used as supporting material to explain the results obtained by means of numeric sensitivity techniques, but also to prevent the use of inadequate techniques. Three-dimensional scatter-plots or XYZ plots show the projection of the sample points (X,Y) on the (Xi,Xj,Y) space. The information they are able to provide is also valuable. The extraction of such information is limited, though challenging, due to obvious interpretation problems when a 3-D figure is shown on a 2-D display. Software packages that allow changing the angle of the view may enhance and broaden their applicability. Extensions of scatter-plots are matrices of scatter-plots and overlay scatter-plots. Matrices of scatter-plots show simultaneously, under a matrix format, the scatter-plots of different pairs of input parameters/output variables. They allow identifying quite quickly the pairs with most remarkable relations, but they are also affected by the loss of accuracy due to including several plots in a reduced space, typically a fraction of a page. Overlay scatter-plots allow showing on the same plot the scatter-plot of one output and several inputs. In order to distinguish the points corresponding to different inputs, different symbols (dots, circles, crosses, diamonds, etc.) and different colours are used. Frequently only a few inputs may be represented due to either the different scales used in the plot or to the difficulties to interpret correctly so many overlapped different symbols. Cobweb plots have been designed to show multidimensional samples in a two-dimensional graph, see Cooke and Van Noortwijk (1999). Vertical parallel lines separated by equal distances are used to represent the sampled values of a given number of inputs/outputs, usually not more than ten or twelve, in order to keep the plot sufficiently clear. Each vertical line is used for a different input/output and either the raw values or the ranks may be represented (either raw values or ranks in all lines, never mixed). Sampled values are marked in each vertical line and jagged lines connect the values corresponding to the same run. Coloured lines can be used to display the different regions of any input parameter or output variable. Moreover, flexible conditioning capabilities enable an extensive insight into particular regions of the mapping. The cobweb plots are usually provided with ‘cross densities’ showing the density of line crossings midway between the vertical axes. Therefore, an informed and careful analysis of cobweb plots enables the characterisation of dependence and conditional dependence. The Contribution to the sample mean plot (CSM plot) represents the fraction of a given output variable mean that is due to any given fraction of smallest values of any input. This is obtained by putting on the x axis the values of the Empirical Cumulative Distribution Function (ECDF) of any input parameter (this values are 1/n, 2/n,…, 1 for any sample of size n of any continuous random variable), and putting on the y axis the fraction of the output variable sample mean corresponding to the smallest value of the input parameter, to the two smallest values, and so on. This way, a monotonic non-decreasing curve is obtained. Plotting the ECDF in the x axis means that equal lengths represent approximately regions of equal probability of the input parameter. The more the plot deviates from the diagonal in a given region, the more (or the less) that region of the input variable contributes to the sample mean or the sample variance. In fact, non-important input variables pro- 392
duce plots close to the diagonal, since large and small output values can be equally found in any of their regions. Additionally, since the values represented on the x axis are independent of the input parameter, many such curves may be plotted in the same graphic corresponding to one output variable and many input parameters. Additionally, a test has been developed to check if deviations from the diagonal are statistically significant or come from pure statistical randomness, see Bolado et al. (2008). Figure 1 is an example of a CSM plot for one output and nine inputs (model used in step 2 of the benchmark on SA techniques). Only inputs W and V (1) show statistically significant deviations from the diagonal. Small values of W are related to large values of the output while the largest values of the output are related to intermediate values of V (1) . For all the other inputs deviations from the diagonal are not statistically significant (large and small values of the output may be obtained in any region of those inputs). Figure 1: Contribution to the sample mean plot for the dose due to 129 I at 9·10 4 y (output variable) and all the input parameters in the model used in the second step of the benchmark on SA techniques. Other graphic tools, as for example the radar plots or the tornado bars, may be used to visualize, in a comparative manner, the value of a given sensitivity index (for example the Pearson correlation coefficient) for a given output and all or a fraction of the input parameters. But these are completely different tools for they show sensitivity indices instead of a representation of the sampled values. 393
- Page 358 and 359: 342
- Page 360 and 361: The Ruprechtov site, located in the
- Page 362 and 363: Colloid Concentration / μg/l 10000
- Page 364 and 365: exists as a stable mineral phase in
- Page 366 and 367: pared to other sites with SOC-beari
- Page 368 and 369: [3] Hauser, W. Geckeis, H., Götz,
- Page 370 and 371: The science and technology group, r
- Page 372 and 373: FEPCAT RTDC 1 RTDC 2 RTDC 3 Clay-ri
- Page 374 and 375: A1: Transport mechanisms Diffusivit
- Page 376 and 377: 360
- Page 378 and 379: maintaining and develop competence
- Page 380 and 381: A project would be justified to det
- Page 382 and 383: 366
- Page 384 and 385: 368
- Page 386 and 387: The main goal of RTDC-1 is to provi
- Page 388 and 389: 3. RTDC3 In RTD component 3 methodo
- Page 390 and 391: lower depths are less saline. For t
- Page 392 and 393: [3] Marivoet, J., Beuth, T., Alonso
- Page 394 and 395: certainty, conducted in RTDC-1 as W
- Page 396 and 397: A simplistic summary might place PA
- Page 398 and 399: There are at least three non-numeri
- Page 400 and 401: 10-11 June 2008. The workshop was a
- Page 402 and 403: Close dialogue between a regulator
- Page 404 and 405: ony, interactions, etc., and to che
- Page 406 and 407: specific sampling strategy, with th
- Page 410 and 411: 3. The sensitivity analysis benchma
- Page 412 and 413: 4. Discussion and conclusions Three
- Page 414 and 415: 398
- Page 416 and 417: 2. Methodology The project particip
- Page 418 and 419: Closely related to this proposal on
- Page 420 and 421: 4.3 Structure The TP structure must
- Page 422 and 423: 4.5 Implementation It is proposed t
- Page 424 and 425: 5.1 The CARD Project has shown that
- Page 426 and 427: � What steps should be taken to m
- Page 428 and 429: 412
- Page 430 and 431: 414
- Page 432 and 433: materials. For attaining the stated
- Page 434 and 435: the bottom of the heated press-mold
- Page 436 and 437: 420
- Page 438 and 439: 422
- Page 440 and 441: focuses on the study of the combine
- Page 442 and 443: emitting radioactive waste to study
- Page 444 and 445: 428
- Page 446 and 447: 2.1 Laboratory experiment The dispo
- Page 448 and 449: 3. Results 3.1 Laboratory experimen
- Page 450 and 451: Barrier. Clays in Natural & Enginee
- Page 452 and 453: 2. Experimental data 2.1 Laboratory
- Page 454 and 455: sented in the accompanying poster,
- Page 456 and 457: 440
duce plots close to the diagonal, since large and small output values can be equally found in any of<br />
their regions. Additionally, since the values represented on the x axis are independent of the input<br />
parameter, many such curves may be plotted in the same graphic corresponding to one output variable<br />
and many input parameters. Additionally, a test has been developed to check if deviations from<br />
the diagonal are statistically significant or come from pure statistical randomness, see Bolado et al.<br />
(2008). Figure 1 is an example of a CSM plot for one output and nine inputs (model used in step 2<br />
of the benchmark on SA techniques). Only inputs W and V (1) show statistically significant deviations<br />
from the diagonal. Small values of W are related to large values of the output while the largest<br />
values of the output are related to intermediate values of V (1) . For all the other inputs deviations<br />
from the diagonal are not statistically significant (large and small values of the output may be obtained<br />
in any region of those inputs).<br />
Figure 1: Contribution to the sample mean plot for the dose due to 129 I at 9·10 4 y (output variable)<br />
and all the input parameters in the model used in the second step of the benchmark on SA techniques.<br />
Other graphic tools, as for example the radar plots or the tornado bars, may be used to visualize, in<br />
a comparative manner, the value of a given sensitivity index (for example the Pearson correlation<br />
coefficient) for a given output and all or a fraction of the input parameters. But these are completely<br />
different tools for they show sensitivity indices instead of a representation of the sampled values.<br />
393