Ross Farrelly

Now up in top 50 of " Don't Get Kicked " by using  Generalized Boosted Regression Models

Am experimenting with R package glmnet in the "Don't Get Kicked" contest on Kaggle .

Just moved into the top 100 in the “Don’t Get Kicked” competition on Kaggle (with a gini of 0.226 vs 1st place of 0.26). Added 0.007 to my Gini score and went up 26 place! It’s a crowded leader board.

Letter to the editor - published in SMH 09/08/2011 The Strategic Review of Indigenous Expenditure confirms there is very little to show for the $3.5 billion a year spent on indigenous programs (''Billions spent but Aborigines little better off, says report'', August 8). This is a travesty for both the intended recipients of the programs and for the taxpayer. It's time for governments to admit they don't know what will work when it comes to improving the lot of indigenous Australians and start to use randomised trials of selected social policies to test what works and what does not. Until this happens very little will change but much will continue to be wasted.

I recently gave a webinar on how food manufacturers in NZ and Australia can comply with the newly introduce Average Quantity System (AQS) legislation. These (currently optional) guidelines indicate how purveyors of pre-packaged good must verify the amount of product contained in the packages they sell. The big challenge is to minimise waste while still maintaining a high probability of passing an audit (0.99 say). As well as covering standard quality techniques such as Gage R&R and Control Charts, I outlined the use of the binomial distribution and the inverse normal distribution to calculate the mean of a process in order to pass an AQS audit with a given probability. To demonstrate the relationship between reducing variation and minimising waste I produced this animation . The area under the curve between 470g and 485g is keep constant, while the standard deviation is varied from 20 down to 3. At each stage the waste, a new mean is calculated and the area under the curve greate...

Last night I attended a talk by Robert Putnam (of “Bowling Alone” fame). He spoke about his new book, “ American Grace ” and the influence of religion on American politics. His main point is that Americans are highly divided on religious matters but also highly tolerant of each others religious views. In brief, Putnam’s thesis is that during the sixties young Americans became less religious but that older Americans became more religious as a reaction to the changes in social mores which took place at the time. This lead to the religious division we see in American today and the rise of the “young nones” – young people who do not identify with any formal religious group. Putnam measures religiosity using survey questions such as “How often do you say grace or give thanks to God before a meal?” However, despite this divide, Americans are very tolerant of people who do hold divergent religious views. Surprisingly, a majority of evangelical Christians in America believe that good peop...

I have been playing around with making the Naplan results more accessible to people who want to examine the data for themselves rather than rely on the summaries published by ACARA. You can find the data accessible via Google maps here: http://www.datamilk.com/schools / where you can create maps such as the one above. Data tables can be found here: http://www.datamilk.com/schools/tables.php and a bit of background info here: http://www.datamilk.com/schools/about.php All comments are welcome.