LaTeX Tutorials

Decent LaTeX lectures here, 1, 2, 3, 4, 5, 6, 7.

Misspecified Model

Could not find the definition of 'misspecified model' directly. However, there are 2 links talking about it -

the statistical model he used to reach his conclusions is "misspecified." This means, in part, that he did not adequately account for other factors which have an impact on crime rates - and which provide an alternate explanations for his findings. When a statistical model is misspecified, it cannot be used as the basis from which to draw conclusions about the impact of policy decisions. One clue that a model is misspecified is if it produces implausible findings.

now if you see that this is being said in a debate against fire arms, you'll take the implausable findings with a pinch (no more) of salt.

Even a misspecified model can be highly accurate in its predictions. Its problems will show up in other ways (e.g. non-random error, which I am guessing is probably the case here; they are probably more likely to be wrong for some cases than for others, e.g. self-funders).
(In point of fact -- most media types tend to make predictions based upon isspecified models. And the biggest problem is not misspecification, but that they do not realize that they are actually using models in the first place. This is one of the many problems that occur when English majors do political science.)

ah that too is from a debate! the other side - lovely!!

oh and from
@inbook { White2006 ,chapter = "Approximate nonlinear forecasting methods",title = "Handbook of Economic Forecasting",volume = "1",author = "Halbert White",publisher = "Elsevier B.V."year = 2006}

When one's goal is to make predictions, the use of a misspecified model is by no means fatal. Our predictions will not be as good as tehy would be if \mu
(true function) were accessable.

Regression Coefficient

An asymmetric measure of association; a statistic computed as part of a regression analysis.

www.ojp.usdoj.gov/BJA/evaluation/glossary/glossary_r.htm

when the regression line is linear (y = ax + b) the regression coefficient is the constant (a) that represents the rate of change of one variable (y) as a function of changes in the other (x); it is the slope of the regression line

wordnet.princeton.edu/perl/webwn

Time-Series Analysis. You can use regression analysis to analyze trends that appear to be related to time.

general knowledge isnt it?

The type of inference

In any kind of choice between techniques, it is important to know the type of inference we want to make. There is no universal solution, because there is no loss-less generalised answer.

Reliability in statistical techniques

Reliability addresses the question of whether repeated application of a procedure will produce similar results.

J Scott Armstron, Fred Collopy,
Error measures for generalising about forecasting methods: Emperical comparisons;
Pg 69-80; International Journal of Forecasting; Vol 8; Year 1992

From before:

Stability is consistency of results, during validation phase, with different samples of data

(Monica Adya and Fred Collopy, J Forecast. 17 481-495 (1998) )

To look at stability in SA, we will define stability as
the consistency of (SA) results within candidate (networks) solutions;

this can be easily justified for SA results in FFNN because we already know that there is high redundancy in free parameters in FFNNs. Therefore, the SA technique that shows consistency between all the networks - and also shows corresponding change in consistency when the data quality is seen to change... promises to be a better technique???