Computation of scores

Fig. 15.14 The upper part shows the computation of scores for a new compound. The lower part shows the prediction of descriptors from the scores.

Barber PA, Demchuk AM, Zhang J, Buchan AM. Validity and reliability of a quantitative computed tomography score in predicting outcome of hyperacute stroke before thrombolytic therapy. ASPECTS Study Group. Alberta Stroke Programme Early CT Score. Lancet 2000 355 1670-1674. 

In the case of a = 1 we compute the scores S for the wind directions in the current example of concentrations of trace elements in atmospheric samples as follows ... 

Then we compute the score of the new observation x on the first PLS dimension and from that we calculate an updated prediction (yj) and we remove the first dimension from eg giving e ... 

Let us look at the typical case in more detail. Suppose we have k dichotomous indicators. The basic strategy of SSMAXCOV is to take a pair (any pair) of indicators as output variables and combine the remaining (k - 2) indicators in one scale. In other words, the input variable is the sum of scores on all indicators, with the exception of the two that were taken as output variables. Then, covariances of the output indicators are computed and plotted for each... 

The NIPALS algorithm is efficient if only a few PCA components are required. Because the deflation procedure increases the uncertainty of following components, the algorithm is not recommended for the computation of many components (Seasholtz et al. 1990). The algorithm fails if convergence is reached already after one cycle in this case another initial value of the score vector has to be tried (Miyashita et al. 1990). 

PCR is an alternative method to the much more used regression method PLS (Section 4.7). PCR is a strictly defined method and the model often gives a very similar performance as a PLS model. Usually PCR needs more components than PLS because no information ofy is used for the computation of the PCA scores this is not necessarily a disadvantage because more variance of X is considered and the model may gain stability. 

Since the goal of SIMCA is to classify a new object jc, a measure for the closeness of the object to the groups needs to be defined. For this purpose, several proposals have been made in the literature. They are based on the orthogonal distance, which represents the Euclidean distance of an object to the PCA space (see Section 3.7.3). First we need to compute the score vector tj of x in theyth PCA space, and using Equation 5.19 and the group center Xj we obtain... 

One could make the task even more difficult by introducing a lag into the joystick correction. This made it much like steering a small boat with a manual rudder, a task tending to produce overshoot. The device automatically computed the score by measuring the area of deviation under a one minute tracing. Practice on this task resulted in continually improving scores, so eventually we eliminated it from the protocol. 

Bohm, H.-J. and Stahl, M. The use of scoring functions in drug discovery applications. Rev. Comput. Chem. 2002, 38,41-87. 

As in the computation of all the scores, one needs to check to make sure that the rank-ordered compounds are reasonable. In the early stages of development of this prioritization method there was a great deal of adjustment to the scoring functions to match medicinal chemists and project team members opinions. 

Having computed the set of scores for all the ECC words, choose the code word b that maximizes the score fonction and output it as the estimated sequence b. ... 

---