# Statistics

Statistics Blogs

## Tests basés sur la vraisemblance – score

October 23, 2014
By
$\widehat\theta_n$

Une autre grandeur intéressant est le score, qui est la dérivée de la vraisemblance. Intuitivement (c’est l’idée de la condition du premier ordre),   et  seront proches si les dérivées en ces points sont proches. En  la dérivée est nulle, donc on va se demander ici, tout simplement, si la dérivée en  est proche de 0. Ou pas. On appele ce test le test du score, ou encore le test du multiplicateur de Lagrange. Ou…

## Tests basés sur la vraisemblance – Rapport de Vraisemblance

October 23, 2014
By
$\widehat\theta_n$

Chose promise, chose due. J’avais dit qu’on parlerait du test de rapport de vraisemblance. L’idée – visuelle – est d’avoir une lecture dans l’autre sens : au lieu de se demander si  et  sont proches, on va se demander si   et  sont proches. Si la fonction de vraisemblance est suffisamment régulière, on se pose la même question. Lorsque j’avais présenté le test, en cours, hier matin, j’avais proposé d’utiliser la delta-method pour…

## Kathryn Chaloner 1954-2014

October 22, 2014
By

Prescript: Memory is fickle. It's been a while since these events, a while since I took her regression course and a while since I've read her papers. One thing I've found is that memories of the contents of particular papers evolves with time, and memo...

## Les tests et la logique, modus tollens

October 22, 2014
By
$A \Rightarrow B$

Quand on apprend la logique, on apprend la notion de modus tollens, correspondant à une notion de contraposition. Si on a une proposition du genre , alors la proposition contraposée est .  Et on apprend que les deux propositions sont équivalentes (je fais de la logique classique). Par exemple si  correspond à “feu” et  à “fumée“,  signifie que tout feu fait de la fumée. Si cette affirmation est vrai, alors il n’y a pas de…

## Prediction intervals too narrow

October 22, 2014
By

Almost all prediction intervals from time series models are too narrow. This is a well-known phenomenon and arises because they do not account for all sources of uncertainty. In my 2002 IJF paper, we measured the size of the problem by computing the actual coverage percentage of the prediction intervals on hold-out samples. We found […]

## Modeling Plenitude and Speciation by Jointly Segmenting Consumers and their Preferences

October 21, 2014
By

In 1993, when music was sold in retail stores, it may have been informative to ask about preference across a handful of music genre. Today, now that the consumer has seized control and the music industry has responded, the market has exploded into more...

October 21, 2014
By

I guess my first posted rant was the call for papers thing.  Here's a second.Countless times, from me to Chair/Dean xxx at Some Other University: I am happy to help with your evaluation of Professor zzz. This email will serve as my lette...

## Degrees of Choice: A modification of a WSJ graphic

October 20, 2014
By

This post uses ggplot2 and rCharts to modify a grouped bar chart that appeared in an article that appeared online on October 19, 2014 in Wall Street Journal’s site. The article was titled “How to Sell a Liberal-Arts Education”. The code for ...

## Tiny Data, Approximate Bayesian Computation and the Socks of Karl Broman

October 20, 2014
By

Big data is all the rage, but sometimes you don’t have big data. Sometimes you don’t even have average size data. Sometimes you only have eleven unique socks: Karl Broman is here putting forward a very interesting problem. Interesting, not onl...

## hts with regressors

October 20, 2014
By

The hts package for R allows for forecasting hierarchical and grouped time series data. The idea is to generate forecasts for all series at all levels of aggregation without imposing the aggregation constraints, and then to reconcile the forecasts so they satisfy the aggregation constraints. (An introduction to reconciling hierarchical and grouped time series is […]