360° Coverage : Don't Underestimate Big Data

2 Updates

Don't Underestimate Big Data

Apr 15 2014, 9:03pm CDT | by

The techniques of big data have been criticized recently in a post by Tim Harford among others. It’s quite easy to imagine the potential here has been exaggerated, and a core point about the...

Filed under: news

 
 
 

20 weeks ago

Don't Underestimate Big Data

Apr 15 2014, 9:03pm CDT | by

The techniques of big data have been criticized recently in a post by Tim Harford among others. It’s quite easy to imagine the potential here has been exaggerated, and a core point about the importance of theory in a lot of empiricism is undeniable. But I think there is some overreach in the criticisms, and in particular one of the failure examples Tim gives shows the opposite of what he claims. He points to a flu forecast by google that failed:

Four years after the original Nature paper was published, Nature News had sad tidings to convey: the latest flu outbreak had claimed an unexpected victim: Google Flu Trends. After reliably providing a swift and accurate account of flu outbreaks for several winters, the theory-free, data-rich model had lost its nose for where flu was going. Google’s model pointed to a severe outbreak but when the slow-and-steady data from the CDC arrived, they showed that Google’s estimates of the spread of flu-like illnesses were overstated by almost a factor of two.

What happened here is in fact one of the advantages of big data: the ability to reject models due to frequently updating data. Google provided a forecast and the forecast and the model underlying it was unequivocably debunked. What a novelty for the social sciences! Of course big data isn’t alone in having constantly updating forecasts that can be tested against reality, but the focus on forecasting and constantly updating data is one of the advantages it has over many other areas of the social sciences. Yes, google made a model and it was falsified. But keep in mind the famous claim John Ioniaddes that most published research findings are false. Maybe it is most, or maybe it’s only half or a quarter. But with big data if a large percent of your findings are false you find out sooner rather than later.

Indeed it is the forecast driven nature of machine learning that makes it so appealing as a branch of empiricism. Maybe this is something you can’t appreciate if you haven’t actually spent time with a big dataset with lots of variables and seen how potentially easy it is to pick the wrong model, or if you haven’t taken a look at the statistical guts of disagreeing research papers. But this focus on the right model being the one that makes the best out of sample predictions, and the ability to be constantly running out-of-sample predictions just seems so much more tied to reality than most research has to be.

Another thing that I think people miss is there is too much focus on what a really long dataset does to, for example, the probability of spurious correlations. What makes machine learning interesting isn’t just dealing with long dataset, but really wide datasets. Consider the data that Netflix has on it’s customers. This isn’t remarkable because of how many people are in the dataset, but how much you know about them. Yes it’s true what they say about spurious correlations, but if you do an out-of-sample prediction with a spurious correlation the relationship will be falsified quickly no matter what your p-value.

In 2001, the statisticisn Leo Breiman wrote a paper titled “Statistical Modeling: The Two Cultures” where he contrasted the data modeling culture and the algorithmic modeling cultures that compete among statisticians. The data modeling culture used goodness-of-fit tests and residual examination to validate models, while the algorithmic modeling culture focused on out-of-sample predictions. He estimated that 98% of all statisticians fell into the former camp, and only 2% fell into the latter. I’m not sure what the numbers are today, and surely many fall into both camps, but the percent of statisticians who are algorithmic folks and the percent of important problems they are working on has gone up sharply from 2%.

It’s true that misleading p-values are a consequence of a lot of data, but this is a data modeling culture problem, not an algorithmic modeling culture problem.

Now don’t get me wrong, I am for the most part a p-value checking, residual examining, data modeling culture economist. But machine learning and big data are going to get more important not less, and I think social scientists who don’t learn to at least think like the other culture are going to be left behind.

 
Update
2

4 days ago

Khazanah throws MAS RM6b lifeline

Aug 29 2014 5:01pm CDT | Source: Business Times Singapore

August 30, 2014 1:15 AMKHAZANAH Nasional will inject RM6 billion (SS$2.4 billion) over three years to resuscitate loss-making Malaysia Airlines (MAS) under a recovery plan that includes even an Act of Parliament. Other key moves are migrating its operations, assets and liabilities to a new company (NewCo) and slashing the workforce ...
Source: Business Times Singapore   Full article at: Business Times Singapore
 

 
Update
1

5 days ago

MAS posts loss of RM307m for Q2

Aug 28 2014 5:00pm CDT | Source: Business Times Singapore

August 29, 2014 1:13 AMMALAYSIA Airlines (MAS) registered a loss of RM307 million (S$122 million) for the second quarter to end-June, but warned of worse to come in the second half when the "full financial impact of the double tragedies of MH370 and MH17" hits hom ...
Source: Business Times Singapore   Full article at: Business Times Singapore
 

 

Don't miss ...

 

<a href="/latest_stories/all/all/30" rel="author">Forbes</a>
Forbes is among the most trusted resources for the world's business and investment leaders, providing them the uncompromising commentary, concise analysis, relevant tools and real-time reporting they need to succeed at work, profit from investing and have fun with the rewards of winning.

 

blog comments powered by Disqus

Latest stories

Watch your diet to reduce diabetes risk
London, Sep 3 (IANS) What you eat and drink, and not just the number of calories, is a significant factor in diabetes risk, a study showed.
 
 
2016 NASA asteroid mission awaits your messages
Washington, Sep 3 (IANS) NASA has invited the public to submit short messages and images on social media that could be placed in a time capsule aboard a spacecraft to be launched in 2016 for an asteroid study.
 
 
Exercise can stem frequent night urination in men
New York, Sep 3 (IANS) Men who are physically active are at lower risk of nocturia - waking up at night to urinate, says a study.
 
 
Nippon Life keen to hike stake in Reliance Life
Tokyo, Sep 3 (IANS) Nippon Life Insurance of Japan wants to hike its stake in its Indian joint venture Reliance Life Insurance to 49 percent from 26 percent, when India notifies the decision of higher foreign equity cap in the sector, a news report here said.
 
 
 

Latest from the Network

Watch your diet to reduce diabetes risk
London, Sep 3 (IANS) What you eat and drink, and not just the number of calories, is a significant factor in diabetes risk, a study showed. Losing weight may be good but not enough to prevent Type 2 diabetes as...
Read more on Celebrity Balla
 
35 militants killed in Afghanistan
Kabul, Sep 3 (IANS) At least 35 Taliban militants have been killed in fresh military operations across Afghanistan, the interior ministry said Wednesday. The Afghan National Security Forces (ANSF) carried out...
Read more on Politics Balla
 
Vince Vaughn for The Politician?
Vince Vaughn is set to star in 'The Politician'. The 'Wedding Crashers' star is attached to take on the lead role in the upcoming Columbia comedy about a scandal-hit politician, according to The Hollywood Reporter. The...
Read more on Movie Balla
 
910 militants, 82 soldiers killed in North Waziristan
Islamabad, Sep 3 (IANS) At least 910 militants and 82 Pakistani soldiers have been killed in operation Zarb-i-Azb in North Waziristan, the military said Wednesday. The statement issued by the Inter-Services Public...
Read more on Politics Balla
 
'Headscarf instils positive body image in British Muslim women'
London, Sep 3 (IANS) British Muslim women who wear a hijab (headscarf) generally have a more positive body image, are less reliant on media messages about beauty ideals and place less importance on appearance than...
Read more on Celebrity Balla
 
Console games new threat to global climate: Study
New York, Sep 3 (IANS) Console games downloaded over broadband internet are causing more greenhouse gas emissions than Blu-ray discs delivered via retail stores, new research reveals. It is not always true that...
Read more on Celebrity Balla
 
Human greed driving species towards extinction: Study
Washington, Sep 3 (IANS) In a disturbing trend, scientists have reported that species are becoming extinct as much as 1,000 times more frequently - 10 times worse than the previous estimates - owing to human population...
Read more on Politics Balla
 
Kelly Brook's traumatic childhood
Kelly Brook has opened up about ''violent and horrific'' memories from her childhood. The 35-year-old actress has revealed that although she enjoyed some aspects of growing up on a council estate in Kent, South East...
Read more on Celebrity Balla
 
Emmy Rossum not 'easy'
Emmy Rossum insists she's not ''easy'' like her 'Shameless' character. The 27-year-old actress stars as Fiona Gallagher in the US remake of the hit series and she's revealed that although she's had sex on a first date,...
Read more on Celebrity Balla
 
2016 NASA asteroid mission awaits your messages
Washington, Sep 3 (IANS) NASA has invited the public to submit short messages and images on social media that could be placed in a time capsule aboard a spacecraft to be launched in 2016 for an asteroid study. Called...
Read more on Apple Balla