Archive

Archive for the ‘data science’ Category

Skimming statistics papers for the ideas (instead of the complete procedures)

June 2nd, 2014 No comments

Been reading a lot of Gelman, Carlin, Stern, Dunson, Vehtari, Rubin “Bayesian Data Analysis” 3rd edition lately. Overall in the Bayesian framework some ideas (such as regularization, and imputation) are way easier to justify (though calculating some seemingly basic quantities becomes tedious). A big advantage (and weakness) of this formulation is statistics has a much less “shrink wrapped” feeling than the classic frequentist presentations. You feel like the material is being written to peers instead of written to calculators (of the human or mechanical variety). In the Bayesian formulation you don’t feel like you will be yelled at for using 1 tablespoon of sugar when the recipe calls for 3 teaspoons (at least if you live in the United States).

Some other stuff reads differently after this though. Read more…

Trimming the Fat from glm() Models in R

May 30th, 2014 10 comments

One of the attractive aspects of logistic regression models (and linear models in general) is their compactness: the size of the model grows in the number of coefficients, not in the size of the training data. With R, though, glm models are not so concise; we noticed this to our dismay when we tried to automate fitting a moderate number of models (about 500 models, with on the order of 50 coefficients) to data sets of moderate size (several tens of thousands of rows). A workspace save of the models alone was in the tens of gigabytes! How is this possible? We decided to find out.

As many R users know (but often forget), a glm model object carries a copy of its training data by default. You can use the settings y=FALSE and model=FALSE to turn this off.

set.seed(2325235)


# Set up a synthetic classification problem of a given size
# and two variables: one numeric, one categorical
# (two levels).
synthFrame = function(nrows) {
   d = data.frame(xN=rnorm(nrows),
      xC=sample(c('a','b'),size=nrows,replace=TRUE))
   d$y = (d$xN + ifelse(d$xC=='a',0.2,-0.2) + rnorm(nrows))>0.5
   d
}


# first show that model=F and y=F help reduce model size

dTrain = synthFrame(1000)
model1 = glm(y~xN+xC,data=dTrain,family=binomial(link='logit'))
model2 = glm(y~xN+xC,data=dTrain,family=binomial(link='logit'),
             y=FALSE)
model3 = glm(y~xN+xC,data=dTrain,family=binomial(link='logit'),
              y=FALSE, model=FALSE)

#
# Estimate the object's size as the size of its serialization
#
length(serialize(model1, NULL))
# [1] 225251
length(serialize(model2, NULL))
# [1] 206341
length(serialize(model3, NULL))
# [1] 189562

dTest = synthFrame(100)
p1 = predict(model1, newdata=dTest, type='response')
p2 = predict(model2, newdata=dTest, type='response')
p3 = predict(model3, newdata=dTest, type='response')
sum(abs(p1-p2))
# [1] 0
sum(abs(p1-p3))
# [1] 0

Read more…

A clear picture of power and significance in A/B tests

May 3rd, 2014 2 comments

A/B tests are one of the simplest reliable experimental designs.

Controlled experiments embody the best scientific design for establishing a causal relationship between changes and their influence on user-observable behavior.

“Practical guide to controlled experiments on the web: listen to your customers not to the HIPPO” Ron Kohavi, Randal M Henne, and Dan Sommerfield, Proceedings of the 13th ACM SIGKDD international conference on Knowledge discovery and data mining, 2007 pp. 959-967.

The ideas is to test a variation (called “treatment” or “B”) in parallel with continuing to test a baseline (called “control” or “A”) to see if the variation drives a desired effect (increase in revenue, cure of disease, and so on). By running both tests at the same time it is hoped that any confounding or omitted factors are nearly evenly distributed between the two groups and therefore not spoiling results. This is a much safer system of testing than retrospective studies (where we look for features from data already collected).

Interestingly enough the multi-armed bandit alternative to A/B testing (a procedure that introduces online control) is one of the simplest non-trivial Markov decision processes. However, we will limit ourselves to traditional A/B testing for the remainder of this note. Read more…

A bit of the agenda of Practical Data Science with R

May 1st, 2014 No comments

The goal of Zumel/Mount: Practical Data Science with R is to teach, through guided practice, the skills of a data scientist. We define a data scientist as the person who organizes client input, data, infrastructure, statistics, mathematics and machine learning to deploy useful predictive models into production.

Our plan to teach is to:

  • Order the material by what is expected from the data scientist.
  • Emphasize the already available bread and butter machine learning algorithms that most often work.
  • Provide a large set of worked examples.
  • Expose the reader to a number of realistic data sets.

Some of these choices may put-off some potential readers. But it is our goal to try and spend out time on what a data scientist needs to do. Our point: the data scientist is responsible for end to end results, which is not always entirely fun. If you want to specialize in machine learning algorithms or only big data infrastructure, that is a fine goal. However, the job of the data scientist is to understand and orchestrate all of the steps (working with domain experts, curating data, using data tools, and applying machine learning and statistics).

Once you define what a data scientist does, you find fewer people want to work as one.

We expand a few of our points below. Read more…

Bandit Formulations for A/B Tests: Some Intuition

April 24th, 2014 2 comments

Controlled experiments embody the best scientific design for establishing a causal relationship between changes and their influence on user-observable behavior.

– Kohavi, Henne, Sommerfeld, “Practical Guide to Controlled Experiments on the Web” (2007)

A/B tests are one of the simplest ways of running controlled experiments to evaluate the efficacy of a proposed improvement (a new medicine, compared to an old one; a promotional campaign; a change to a website). To run an A/B test, you split your population into a control group (let’s call them “A”) and a treatment group (“B”). The A group gets the “old” protocol, the B group gets the proposed improvement, and you collect data on the outcome that you are trying to achieve: the rate that patients are cured; the amount of money customers spend; the rate at which people who come to your website actually complete a transaction. In the traditional formulation of A/B tests, you measure the outcomes for the A and B groups, determine which is better (if either), and whether or not the difference observed is statistically significant. This leads to questions of test size: how big a population do you need to get reliably detect a difference to the desired statistical significance? And to answer that question, you need to know how big a difference (effect size) matters to you.

The irony is that to detect small differences accurately you need a larger population size, even though in many cases, if the difference is small, picking the wrong answer matters less. It can be easy to lose sight of that observation in the struggle to determine correct experiment sizes.

There is an alternative formulation for A/B tests that is especially suitable for online situations, and that explicitly takes the above observation into account: the so-called multi-armed bandit problem. Imagine that you are in a casino, faced with K slot machines (which used to be called “one-armed bandits” because they had a lever that you pulled to play (the “arm”) — and they pretty much rob you of all your money). Each of the slot machines pays off at a different (unknown) rate. You want to figure out which of the machines pays off at the highest rate, then switch to that one — but you don’t want to lose too much money to the suboptimal slot machines while doing so. What’s the best strategy?

NewImage

The “pulling one lever at a time” formulation isn’t a bad way of thinking about online transactions (as opposed to drug trials); you can imagine all your customers arriving at your site sequentially, and being sent to bandit A or bandit B according to some strategy. Note also, that if the best bandit and the second-best bandit have very similar payoff rates, then settling on the second best bandit, while not optimal, isn’t necessarily that bad a strategy. You lose winnings — but not much.

Traditionally, bandit games are infinitely long, so analysis of bandit strategies is asymptotic. The idea is that you test less as the game continues — but the testing stage can go on for a very long time (often interleaved with periods of pure exploitation, or playing the best bandit). This infinite-game assumption isn’t always tenable for A/B tests — for one thing, the world changes; for another, testing is not necessarily without cost. We’ll look at finite games below.

Read more…

Can a classifier that never says “yes” be useful?

March 8th, 2014 2 comments

Many data science projects and presentations are needlessly derailed by not having set shared business relevant quantitative expectations early on (for some advice see Setting expectations in data science projects). One of the most common issues is the common layman expectation of “perfect prediction” from classification projects. It is important to set expectations correctly so your partners know what you are actually working towards and do not consider late choices of criteria disappointments or “venue shopping.” Read more…

Drowning in insignificance

February 26th, 2014 5 comments

Some researchers (in both science and marketing) abuse a slavish view of p-values to try and falsely claim credibility. The incantation is: “we achieved p = x (with x ≤ 0.05) so you should trust our work.” This might be true if the published result had been performed as a single project (and not as the sole shared result in longer series of private experiments) and really points to the fact that even frequentist significance is a subjective and intensional quantity (an accusation usually reserved for Bayesian inference). In this article we will comment briefly on the negative effect of un-reported repeated experiments and what should be done to compensate. Read more…

The gap between data mining and predictive models

February 20th, 2014 3 comments

The Facebook data science blog shared some fun data explorations this Valentine’s Day in Carlos Greg Diuk’s “The Formation of Love”. They are rightly receiving positive interest in and positive reviews of their work (for example Robinson Meyer’s Atlantic article). The finding is also a great opportunity to discuss the gap between cool data mining results and usable predictive models. Data mining results like this (and the infamous “Beer and Diapers story”) face an expectation that one is immediately ready to implement something like what is claimed in: “Target Figured Out A Teen Girl Was Pregnant Before Her Father Did” once an association is plotted.

Producing a revenue improving predictive model is much harder than mining an interesting association. And this is what we will discuss here. Read more…

Unprincipled Component Analysis

February 10th, 2014 4 comments

As a data scientist I have seen variations of principal component analysis and factor analysis so often blindly misapplied and abused that I have come to think of the technique as unprincipled component analysis. PCA is a good technique often used to reduce sensitivity to overfitting. But this stated design intent leads many to (falsely) believe that any claimed use of PCA prevents overfit (which is not always the case). In this note we comment on the intent of PCA like techniques, common abuses and other options.

The idea is to illustrate what can quietly go wrong in an analysis and what tests to perform to make sure you see the issue. The main point is some analysis issues can not be fixed without going out and getting more domain knowledge, more variables or more data. You can’t always be sure that you have insufficient data in your analysis (there is always a worry that some clever technique will make the current data work), but it must be something you are prepared to consider. Read more…

Bad Bayes: an example of why you need hold-out testing

February 1st, 2014 3 comments

We demonstrate a dataset that causes many good machine learning algorithms to horribly overfit.

The example is designed to imitate a common situation found in predictive analytic natural language processing. In this type of application you are often building a model using many rare text features. The rare text features are often nearly unique k-grams and the model can be anything from Naive Bayes to conditional random fields. This sort of modeling situation exposes the modeler to a lot of training bias. You can get models that look good on training data even though they have no actual value on new data (very poor generalization performance). In this sort of situation you are very vulnerable to having fit mere noise.

Often there is a feeling if a model is doing really well on training data then must be some way to bound generalization error and at least get useful performance on new test and production data. This is, of course, false as we will demonstrate by building deliberately useless features that allow various models to perform well on training data. What is actually happening is you are working through variations of worthless models that only appear to be good on training data due to overfitting. And the more “tweaking, tuning, and fixing” you try only appears to improve things because as you peek at your test-data (which you really should have held some out until the entire end of project for final acceptance) your test data is becoming less exchangeable with future new data and more exchangeable with your training data (and thus less helpful in detecting overfit).

Any researcher that does not have proper per-feature significance checks or hold-out testing procedures will be fooled into promoting faulty models. Read more…