Inspection, graphical

Data should be subjected to graphical inspection for any strong evidence of nonlinearity before launching into a formal correlation analysis. 

The results are presented in Figure 3.9. Underlying trends are not obvious from inspection of the raw data. Of course, further madiematical analysis might reveal a systematic trend, but in most situations the first inspection is graphical. In real time situations, such as process control, on-line graphical inspection is essential. The five point MA... 

Graphical inspection of response surfaces is restricted to three-dimensional plots. How do you plot response surfaces if more than two factors are included in the study ... 

For graphical inspection of regression models, the analysis of residuals is apphcable. The residual, e, denotes the difference between the observed value, y, and the value estimated by the model, y . For n observations, we get... 

Displaying the drinks recipes in this manner makes interpretation of the data significantly easier. It is possible now, simply by graphical inspection, to determine whether any drink of interest is achievable or not. Let us consider each drink in turn and discuss why it may or may not be physically realizable given the mixing constraints ... 

We shall be interested in plotting a PFR trajectory from the feed conditions, specified in Chapter 1, in Cb-Ce-Ct space. This example also wishes to demonstrate how integration limits, for the PFR equation, may be estimated from graphical inspection of the species concentration profiles. 

A process or procedure, such as ultrasonic or radio-graphic inspection, for determining the quality or characteristics of a material, part, or assembly, without permanently altering the subject or its properties. Used to find internal anomalies in a structure without degrading its properties. 

Impedance measurements produce numerical results, usually as real Z and imaginary Z" impedances or modulus IZl and phase angle q> as functions of frequency. Visual (graphical) inspection of the obtained results usually makes it possible to identify the electrical equivalent circuit containing R, C, and L elements. 

However, this inspection is insufficient, and mathematical modeling involving fit to the circuit or equation should be carried out (Chap. 14). In the case of real electrochemical systems, the situation is more complex because the studied objects are not electrical circuits but systems involving interfaces, electrochemical reactions, transport of species, etc. Nevertheless, graphical inspection usually helps in deciding whether the experiments are proceeding correctly and in making a first assessment of data. 

Graphical inspection of loadings Pa, fl = 1,2,..., A was also used as an extra validation factors that primarily reflect random noise are ignored in the final model. 

Compared to PCR, PLSR usually gives more of the relevant modeling in the first few factors, making the graphical inspection of the model easier. 

Compared to, say, SMLR the bilinear PLSR calibration model is usually easier to understand The SMLR (and PLSR) regression coefficient vector b is a contrast between all the different phenomena affecting the NIR spectra, so it can be very confusing. The bilinear PLSR allows us to study these NIR phenomena more or less individually, by graphical inspection of the A-dimensional model subspace. 

When, as illustrated here, PCA is used as an EMDA tool, as the purpose is graphical representati(Mi/inspecti(Mi of data, the matter of choosing an appropriate number of PCs, at first sight, does not seem so relevant. This is tree, and not true, at the same time. True because it is always possible to calculate all components up to the rank and idenrify the most significant PCs by sequential graphical inspection of score plots. Not fine because the residual E (the errors part or the not systematic variatiOTi in data) does cmistitute a relevant part of what we also want to know about our data, such as outliers, noise content. More generally, we may see... 

---