If I have an AB testing result, and the general A/B lift is small (target rate 5% vs 5.1%). How can I find segments/subsets of the population that have higher A/B lift using other features?
For example, the dataset has binary dependent variable y=0 or 1, independent features x1-x10, and treatment A and B. The test result shows A has 5% target rate, and B has 5.1%. If I want to find subsets of the dataset which have larger target rate difference between A and B by using the features x1-x10, What methods can I use? (What kind of algorithm and how to design the dependent variable in this case?)
I'm thinking about building a customized decision tree that instead of splitting on the optimized entropy/gini, it splits on the highest A% and B% difference (maybe like (A-B)log(A/B)). Haven't tried yet, just want to know any standard ways to handle it.
Related
Given n samples with d features of stock A, we can build a (d+1) dimensional linear model to predict the profit. However, in some books, I found that if we have m different stocks with n samples and d features for each, then they merge these data to get m*n samples with d features to build a single (d+1) dimensional linear model to predict the profit.
My confusion is that, different stocks usually have little connection with each other, and their profit are influenced by different factors and environment, so why they can be merged to build a single model?
If you are using R as tool of choice, you might like the time series embedding howto and its appendix -- the mathematics behind that is Taken's theorem:
[Takens's theorem gives] conditions under which a chaotic dynamical system can be reconstructed from a sequence of observations of the state of a dynamical system.
It looks to me as the statement's you quote seem to relate to exactly this theorem: For d features (we are lucky, if we know that number - we usually don't), we need d+1 dimensions.
If more time series should be predicted, we can use the same embedding space if the features are the same. The dimensions d are usually simple variables (like e.g. temperature for different energy commodity stocks) - this example helped me to intuitively grasp the idea.
Further reading
Forecasting with Embeddings
Lets say I have 100 independent features - 90 are binary (e.g. 0/1) and 10 are continuous variables (e.g. age, height, weight, etc). I use the 100 features to predict a classifier problem with an adequate amount of samples.
When I set a XGBClassifier function and fit it, then the 10 most important features from the standpoint of gain are always the 10 continuous variable. For now I am not interested in cover or frequency. The 10 continuous variables take up like .8 to .9 of space in gain list ( sum(gain) = 1).
I tried tuning the gamma, reg_alpha , reg_lambda , max_depth, colsample. Still top 10 features by gain are always the 10 continuous features.
Any suggestions?
small update -- someone asked why I think this is happening. I believe it's because a continuous variable can be split on multiple times per decision tree. A binary variable can only be split on once. Hence, the higher prevalence of continuous variables in trees and thus a higher gain score
Yes, it's well-known that a tree(/forest) algorithm (xgboost/rpart/etc.) will generally 'prefer' continuous variables over binary categorical ones in its variable selection, since it can choose the continuous split-point wherever it wants to maximize the information gain (and can freely choose different split-points for that same variable at other nodes, or in other trees). If that's the optimal tree (for those particular variables), well then it's the optimal tree. See Why do Decision Trees/rpart prefer to choose continuous over categorical variables? on sister site CrossValidated.
When you say "any suggestions", depends what exactly do you want, it could be one of the following:
a) To find which of the other 90 binary categorical features give the most information gain
b) To train a suboptimal tree just to find out which features those are
c) To engineer some "compound" features by combining the binary features into n-bit categorical features which have more information gain (while being sure to remove the individual binary features from the input)
d) You could look into association rules : What is the practical difference between association rules and decision trees in data mining?
If you want to explore a)...c), suggest something vaguely like this:
exclude various subsets of the 10 continuous variables, then see which binary features show up as having the most gain. Let's say that gives you N candidate features. N will be << 90, let's assume N < 20 to make the following more computationally efficient.
then compute the pairwise measure of association or correlation (Spearman or Kendall) between each of the N features. Look at a corrplot. Pick the clusters of variables which are most associated with each other. Create compound n-bit variables which combine those individual binary features. Then retrain the tree, including the compound variables, and excluding the individual binary variables (to avoid changing the total variance in the input).
iterate for excluding various subsets of the 10 continuous variables. See which patterns emerge in your compound variables. I'm sure there's an algorithm for doing this (compound feature-engineering of n-bit categoricals) more formally and methodically, I just don't know it.
Anyway, for hacking a tree-based method for better performance, I imagine the most naive way is "at every step, pick the two most highly-correlated/associated categorical features and combine them". Then retrain the tree (include new feature, exclude its constituent features) and use the revised gain numbers.
perhaps a more robust way might be:
Pick some threshold T for correlation/association, say start at a high level T = 0.9 or 0.95
At each step, merge any features whose absolute correlation/association to each other >= T
If there were no merges at this step, reduce T by some value (like T -= 0.05) or ratio (e.g. T *= 0.9 . If still no merges, keep reducing T until there are merges, or until you hit some termination value (e.g. T = 0.03)
Retrain the tree including the compound variables, excluding their constituent subvariables.
Now go back and retrain what should be an improved tree with all 10 continuous variables, and your compound categorical features.
Or you could early-terminate the compound feature selection to see what the full retrained tree looks like.
This issue arose in the 2014 Kaggle Allstate Purchase Prediction Challenge, where the policy coverage options A,B,C,D,E,F,G were each categoricals with between 2-4 values, and very highly correlated with each other. (The current option of C, "C_previous", is one of the input features). See that competitions's forums and published solutions for more. Be aware that policy = (A,B,C,D,E,F,G) is the output. But C_previous is an input variable.
Some general fast-and-dirty rules-of-thumb on feature selection from Kaggle are:
throw out any near-constant/ very-low-variance variables (because they have near-zero information content)
throw out any very-high-cardinality categorical variables (cardinality >~ training-set-size/2), (because they will also tend to have low information content, but cause lots of spurious overfitting and blow up training time). This can include customer IDs, row IDs, transaction IDs, sequence IDs, and other variables which shouldn't be trained on in the first place but accidentally ended up in the training set.
I can suggest few things for you to try.
Test your model without this data (only 90 features) and evaluate the decrease in your score. If it's insignificant you might want to remove those features.
Turn them into groups.
For example, age can be categorized into groups, 0 : 0-7, 1 : 8-16, 2 : 17-25 and so on.
Turn them into binary. Out of the box idea on how to chose the best value to split them into binary is: Build 1 tree with 1 node (max depth = 1) and use only 1 feature. (1 out of the continuous features). then, dump the model to a .txt file and see the value it chose to split on. using this value, you can transform all that feature column into binary
I'm dealing myself with very similar problems right now, So i'll be happy to hear your results and the paths you chose to try.
I learned a lot from the answer by #smci, so I would recommend to follow his suggestions.
In the case, when your binary categorical features are in fact OHE representations of several categorical features with several classes in each, you can follow two more approaches:
Convert OHE into label encoding. Yes, this has the caveat that one introduces an order into a categorical features, which might be meaningless, for example green=3 > red=2 > blue=1. But in practice is seems that trees handle label=encoded categorical variables (even with meaningless order) reasonably well.
Convert OHE into target-/mean-/likelihood encoding. This is tricky, because you need to apply regularisation to avoid data leakage.
Both of those ideas are meant to group together several binary features into a single one based on prior knowledge about feature meaning. If you do not have that luxury, you can also try to deduce such groups by doing scalar product of columns and finding those giving zero product.
I want to create a synthetic dataset consisting of 2 classes and 3 features for testing a hyperparameter optimization technique for a SVM classifier with a RBF kernel. The hyperparameters are gamma and C (the cost).
I have created my current 3D synthetic dataset as follows:
I have created 10 based points for each class by sampling from a multivariate normal distribution with mean (1,0,0) and (0,1,0), respectively, and unit variance.
I have added more points to each class by picking a base point at random and then sampling a new point from a normal distribution with mean equal to the chosen base point and variance I/5.
It would be a very cool thing if I could determine the best C and gamma from the dataset (before running SVM), so that I can see if my optimization technique provides me the best parameters in the end.
Is there a possibility to calculate the best gamma and C parameter from the synthetic dataset described above?
Or else is there a way to create a synthetic dataset where the best gamma and C parameters are known?
Very interesting question, but the answer is no. It is completely data specific, even knowing exactly the distributions, unless you have an infinite sample, it is mathematicaly impossible to prove best C/gamma as SVM in the end is purely point-based method (as opposed to density estimation based). Typical comparison is done in a different scenario - you take real data, and fit hyperparams using other techniques, like gaussian processes (bayesian optimization) etc, which generate baseline (and probably will get to optimal C and gamma too, or at least realy close to them). In the end looking for best C and gamma is not complex problem, thus simply run good techniqe (like bayesopt) for a longer time, and you will get your optimas to compare against. Furthermore, remember that the task of hyperparams optimization is not to find a particular C and gamma, it is to find hyperparams yielding best results, and in fact, even for SVM, there might be many sets of "optimal" C and gammas, all yielding the same results (in terms of your finite dataset) despite being very far away from each other.
I need some point of view to know if what I am doing is good or wrong or if there is better way to do it.
I have 10 000 elements. For each of them I have like 500 features.
I am looking to measure the separability between 2 sets of those elements. (I already know those 2 groups I don't try to find them)
For now I am using svm. I train the svm on 2000 of those elements, then I look at how good the score is when I test on the 8000 other elements.
Now I would like to now which features maximize this separation.
My first approach was to test each combination of feature with the svm and follow the score given by the svm. If the score is good those features are relevant to separate those 2 sets of data.
But this takes too much time. 500! possibility.
The second approach was to remove one feature and see how much the score is impacted. If the score changes a lot that feature is relevant. This is faster, but I am not sure if it is right. When there is 500 feature removing just one feature don't change a lot the final score.
Is this a correct way to do it?
Have you tried any other method ? Maybe you can try decision tree or random forest, it would give out your best features based on entropy gain. Can i assume all the features are independent of each other. if not please remove those as well.
Also for Support vectors , you can try to check out this paper:
http://axon.cs.byu.edu/Dan/778/papers/Feature%20Selection/guyon2.pdf
But it's based more on linear SVM.
You can do statistical analysis on the features to get indications of which terms best separate the data. I like Information Gain, but there are others.
I found this paper (Fabrizio Sebastiani, Machine Learning in Automated Text Categorization, ACM Computing Surveys, Vol. 34, No.1, pp.1-47, 2002) to be a good theoretical treatment of text classification, including feature reduction by a variety of methods from the simple (Term Frequency) to the complex (Information-Theoretic).
These functions try to capture the intuition that the best terms for ci are the
ones distributed most differently in the sets of positive and negative examples of
ci. However, interpretations of this principle vary across different functions. For instance, in the experimental sciences χ2 is used to measure how the results of an observation differ (i.e., are independent) from the results expected according to an initial hypothesis (lower values indicate lower dependence). In DR we measure how independent tk and ci are. The terms tk with the lowest value for χ2(tk, ci) are thus the most independent from ci; since we are interested in the terms which are not, we select the terms for which χ2(tk, ci) is highest.
These techniques help you choose terms that are most useful in separating the training documents into the given classes; the terms with the highest predictive value for your problem. The features with the highest Information Gain are likely to best separate your data.
I've been successful using Information Gain for feature reduction and found this paper (Entropy based feature selection for text categorization Largeron, Christine and Moulin, Christophe and Géry, Mathias - SAC - Pages 924-928 2011) to be a very good practical guide.
Here the authors present a simple formulation of entropy-based feature selection that's useful for implementation in code:
Given a term tj and a category ck, ECCD(tj , ck) can be
computed from a contingency table. Let A be the number
of documents in the category containing tj ; B, the number
of documents in the other categories containing tj ; C, the
number of documents of ck which do not contain tj and D,
the number of documents in the other categories which do
not contain tj (with N = A + B + C + D):
Using this contingency table, Information Gain can be estimated by:
This approach is easy to implement and provides very good Information-Theoretic feature reduction.
You needn't use a single technique either; you can combine them. Term-Frequency is simple, but can also be effective. I've combined the Information Gain approach with Term Frequency to do feature selection successfully. You should experiment with your data to see which technique or techniques work most effectively.
If you want a single feature to discriminate your data, use a decision tree, and look at the root node.
SVM by design looks at combinations of all features.
Have you thought about Linear Discriminant Analysis (LDA)?
LDA aims at discovering a linear combination of features that maximizes the separability. The algorithm works by projecting your data in a space where the variance within classes is minimum and the one between classes is maximum.
You can use it reduce the number of dimensions required to classify, and also use it as a linear classifier.
However with this technique you would lose the original features with their meaning, and you may want to avoid that.
If you want more details I found this article to be a good introduction.
I am working on a Machine Learning problem which looks like this:
Input Variables
Categorical
a
b
c
d
Continuous
e
Output Variables
Discrete(Integers)
v
x
y
Continuous
z
The major issue that I am facing is that Output Variables are not totally independent of each other and there is no relation that can be established between them. That is, there is a dependence but not due to the causality (one value being high doesn't imply that the other will be high too but the chances of other being higher will improve)
An Example would be:
v - Number of Ad Impressions
x - Number of Ad Clicks
y - Number of Conversions
z - Revenue
Now, for an Ad to be clicked, it has to first appear on a search, so Click is somewhat dependent on Impression.
Again, for an Ad to be Converted, it has to be first clicked, so again Conversion is somewhat dependent on Click.
So running 4 instances of the problem predicting each of the output variables doesn't make sense to me. Infact there should be some way to predict all 4 together taking care of their implicit dependencies.
But as you can see, there won't be a direct relation, infact there would be a probability that is involved but which can't be worked out manually.
Plus the output variables are not Categorical but are in fact Discrete and Continuous.
Any inputs on how to go about solving this problem. Also guide me to existing implementations for the same and which toolkit to use to quickly implement the solution.
Just a random guess - I think this problem can be targeted by Bayesian Networks. What do you think ?
Bayesian Networks will do fine in your case. Your network won't be that huge either so you can live with exact inference algorithms like graph elimination or junction tree. If you decide to use BNs, then you can use Kevin Murphy's BN toolbox. Here is a link to that. For a more general toolbox that uses Gibbs sampling for approximate Monte Carlo inference, you can use BUGS.
Edit:
As an example look at the famous sprinkler example here. For totally discrete variables, you define the conditional probability tables as in the link. For instance you say that given that today is cloudy, there is a 0.8 probability of rain. You define all probability distributions, where the graph shows the causality relations (i.e. if cloud then rain etc.) Then as query you ask to your inference algorithm questions like, given that grass was wet; was it cloudy, was it raining, was the sprinkler on and so on.
To use BNs one needs a system model that is described in terms of causality relations (Directed Acyclic Graph) and probability transitions. If you wanna learn your system parameters there are techniques like EM algorithm. However, learning the graph structure is a really hard task and supervised machine learning approaches will do better in that case.