Two weeks ago I used STAN to create predictions after just throwing in all independent variables. This week I aim to refine the STAN model. For this it is convenient to use the loo package (Efficient Leave-One-Out Cross-Validation and WAIC for Bayesian... [Read more...]

Two weeks ago I used STAN to create predictions after just throwing in all independent variables. This week I aim to refine the STAN model. For this it is convenient to use the loo package (Efficient Leave-One-Out Cross-Validation and WAIC for Bayesian Models). See also the paper by Aki Vehtari, ... [Read more...]

It is a bit a contradiction. Kaggle provides competitions on data science, while Stan is clearly part of the (Bayesian) statistics. Yet after using random forests, boosting and bagging, I also think this problem has a suitable size for Stan, which I understand can handle larger problems than older Bayesian ... [Read more...]

It is a bit a contradiction. Kaggle provides competitions on data science, while Stan is clearly part of the (Bayesian) statistics. Yet after using random forests, boosting and bagging, I also think this problem has a suitable size for Stan, which I un... [Read more...]

In two previous posts (Predicting Titanic deaths on Kaggle IV: random forest revisited, Predicting Titanic deaths on Kaggle) I was unable to make random forest predict as well as boosting. Hence when I read about an alternative implementation; ranger I took the opportunity to check if with ranger I could ... [Read more...]

In two previous posts (Predicting Titanic deaths on Kaggle IV: random forest revisited, Predicting Titanic deaths on Kaggle) I was unable to make random forest predict as well as boosting. Hence when I read about an alternative implementation; ranger&n... [Read more...]

On July 19th I used randomForest to predict the deaths on Titanic in the Kaggle competition. Subsequently I found that both bagging and boosting gave better predictions than randomForest. This I found somewhat unsatisfactory, hence I am now revisiting randomForest. To my disappointment this does not result in predictions as ... [Read more...]

On July 19th I used randomForest to predict the deaths on Titanic in the Kaggle competition. Subsequently I found that both bagging and boosting gave better predictions than randomForest. This I found somewhat unsatisfactory, hence I am now revisi... [Read more...]

This is the third post on prediction the deaths. The first one used randomforest, the second boosting (gbm). The aim of the third post was to use bagging. In contrast to the former posts I abandoned dplyr in this post. It gave some now you see now you ... [Read more...]

This is the third post on prediction the deaths. The first one used randomforest, the second boosting (gbm). The aim of the third post was to use bagging. In contrast to the former posts I abandoned dplyr in this post. It gave some now you see now you don't errors.... [Read more...]

Following my previous post I have decided to try and use a different method: generalized boosted regression models (gbm). I have read the background in Elements of Statistical Learning and arthur charpentier's nice post on it. This data ... [Read more...]

Following my previous post I have decided to try and use a different method: generalized boosted regression models (gbm). I have read the background in Elements of Statistical Learning and arthur charpentier's nice post on it. This data is a nice occasion to get my hands dirty.Data Data as ... [Read more...]

Kaggle has a competition to predict who will die on the famous Titanic 'Machine Learning from Disaster''. It is placed as knowledge competition. Just up there to learn. I am late to the party, it has been been for 1 1/2 year, to end by end 2015. It is a small data set, ... [Read more...]

Kaggle has a competition to predict who will die on the famous Titanic 'Machine Learning from Disaster''. It is placed as knowledge competition. Just up there to learn. I am late to the party, it has been been for 1 1/2 year, to end by end 2015. It is ... [Read more...]

Last week I had a post 'Deaths in the Netherlands by cause and age'. During creation of that post I made one plot which I had not shown. It shows something odd. There is a vertical striping. Hence mortality varies by year across age.To examine this phe... [Read more...]

I downloaded counts of deaths by age, year and mayor cause from the Dutch statistics site. In this post I do some plots to look at causes and changes between the years.Data Data from CBS. I downloaded the data in Dutch, hence the first thing to do... [Read more...]

I restarted at working my way through the PROC MCMC examples. The SAS manual describes this example: Consider the data set from Bacon and Watts (1971), where  is the logarithm of the height of the stagnant surface layer and the covariate... [Read more...]

I am thinking about a new laptop. For one thing a 1366*768 resolution just seems to get impractically small. Secondly, faster comutations, more memory.Regarding CPU speed, my current laptop has a lowly Celeron 877. From what I see at my computers activ... [Read more...]

I was told the Eurostat package would be interesting for me.  This is indeed true and now I want to use it to plot some data which are related core of some of the European policies; debt.In these plots I only show individual countries, not aggrega... [Read more...]

As final part of my paper helicopter experiment analysis (part I, part II) I do a reanalysis for one more data set. In 2002 Erik Erhardt and Hantao Mai did an extensive experiment, see The Search for the Optimal Paper Helicopter. They did a number of s... [Read more...]