I was trying to compare the SEs from two approaches:
the SE printed in the results section in the GLM poisson model, vs.
the SE of the coefficient I got from bootstrapping (basically I resampled the same number of observations with replacement and fit a new Poisson model and get the coefficient, then I compute the SD of the coefficients from 1000 bootstraps)
However the SEs from the two approaches are quite different. The SE from the 1) is much smaller. I thought they would be numerically similar.
Anyone has a explanation?
Hard to answer without a reproducible example. The usual suspicion would be that there is overdispersion in the response variable, leading to conventional standard errors that are too small. Alternatives include standard errors from: bootstrap, sandwich, quasi-Poisson, or fitting a negative binomial model instead.
See Section 3 of https://doi.org/10.18637/jss.v027.i08 for a worked example in R as well as Chapter 7.4 in https://discdown.org/microeconometrics/. These use the sandwich() function from the package sandwich which also includes a function vcovBS() for bootstrap covariances.
Related
I've a dataset (based on million song dataset) on which I need to do genre classification. Following is the distribution of various genre classes in the dataset.
Genre Count %age
1. Rock 115104 39.94364359
2. Pop 47534 16.49535337
3. Electronic 24313 8.437150809
4. Jazz 16465 5.713720564
5. Rap 15347 5.325749741
6. RnB 13769 4.778148706
7. Country 13509 4.687922933
8. Reggae 8739 3.032627027
9. Blues 7075 2.455182083
10. Latin 7042 2.44373035
11. Metal 6257 2.171317921
12. World 4624 1.604630664
13. Folk 3661 1.270448283
14. Punk 3479 1.207290242
15. New Age 1248 0.433083709
Would you call this data unbalanced? I've tried reading around but found that people describe datasets unbalanced where one of the classes is 99% of the dataset and it's a binary classification problem. Not sure whether the above set falls into this category. Please help. I'm not able to get the classification right and being a newbie can't decide whether it's the data or my naivety. This is one of the hypotheses I have and need to validate.
In general there is no strict definition of imbalanced dataset, but generally, if the smallest class is 10x smaller than the largest one, then calling it imbalanced is a good idea.
In your case, smallest class is actually 100x smaller than the biggest one, so you can even map it onto your consideration of "99-1" for binary classification. If you only ask about differentiating between New Age and Rock you will end up with 99-1 imbalance, so you can expect problems typical to imbalanced classification - to appear in your project.
I have a dataset X whose shape is (1741, 61). Using logistic regression with cross_validation I was getting around 62-65% for each split (cv =5).
I thought that if I made the data quadratic, the accuracy is supposed to increase. However, I'm getting the opposite effect (I'm getting each split of cross_validation to be in the 40's, percentage-wise) So,I'm presuming I'm doing something wrong when trying to make the data quadratic?
Here is the code I'm using,
from sklearn import preprocessing
X_scaled = preprocessing.scale(X)
from sklearn.preprocessing import PolynomialFeatures
poly = PolynomialFeatures(3)
poly_x =poly.fit_transform(X_scaled)
classifier = LogisticRegression(penalty ='l2', max_iter = 200)
from sklearn.cross_validation import cross_val_score
cross_val_score(classifier, poly_x, y, cv=5)
array([ 0.46418338, 0.4269341 , 0.49425287, 0.58908046, 0.60518732])
Which makes me suspect, I'm doing something wrong.
I tried transforming the raw data into quadratic, then using preprocessing.scale, to scale the data, but it was resulting in an error.
UserWarning: Numerical issues were encountered when centering the data and might not be solved. Dataset may contain too large values. You may need to prescale your features.
warnings.warn("Numerical issues were encountered "
So I didn't bother going this route.
The other thing that's bothering is the speed of the quadratic computations. cross_val_score is taking around a couple of hours to output the score when using polynomial features. Is there any way to speed this up? I have an intel i5-6500 CPU with 16 gigs of ram, Windows 7 OS.
Thank you.
Have you tried using the MinMaxScaler instead of the Scaler? Scaler will output values that are both above and below 0, so you will run into a situation where values with a scaled value of -0.1 and those with a value of 0.1 will have the same squared value, despite not really being similar at all. Intuitively this would seem to be something that would lower the score of a polynomial fit. That being said I haven't tested this, it's just my intuition. Furthermore, be careful with Polynomial fits. I suggest reading this answer to "Why use regularization in polynomial regression instead of lowering the degree?". It's a great explanation and will likely introduce you to some new techniques. As an aside #MatthewDrury is an excellent teacher and I recommend reading all of his answers and blog posts.
There is a statement that "the accuracy is supposed to increase" with polynomial features. That is true if the polynomial features brings the model closer to the original data generating process. Polynomial features, especially making every feature interact and polynomial, may move the model further from the data generating process; hence worse results may be appropriate.
By using a 3 degree polynomial in scikit, the X matrix went from (1741, 61) to (1741, 41664), which is significantly more columns than rows.
41k+ columns will take longer to solve. You should be looking at feature selection methods. As Grr says, investigate lowering the polynomial. Try L1, grouped lasso, RFE, Bayesian methods. Try SMEs (subject matter experts who may be able to identify specific features that may be polynomial). Plot the data to see which features may interact or be best in a polynomial.
I have not looked at it for a while but I recall discussions on hierarchically well-formulated models (can you remove x1 but keep the x1 * x2 interaction). That is probably worth investigating if your model behaves best with an ill-formulated hierarchical model.
I have a large dataset I am trying to do cluster analysis on using SOM. The dataset is HUGE (~ billions of records) and I am not sure what should be the number of neurons and the SOM grid size to start with. Any pointers to some material that talks about estimating the number of neurons and grid size would be greatly appreciated.
Thanks!
Quoting from the som_make function documentation of the som toolbox
It uses a heuristic formula of 'munits = 5*dlen^0.54321'. The
'mapsize' argument influences the final number of map units: a 'big'
map has x4 the default number of map units and a 'small' map has
x0.25 the default number of map units.
dlen is the number of records in your dataset
You can also read about the classic WEBSOM which addresses the issue of large datasets
http://www.cs.indiana.edu/~bmarkine/oral/self-organization-of-a.pdf
http://websom.hut.fi/websom/doc/ps/Lagus04Infosci.pdf
Keep in mind that the map size is also a parameter which is also application specific. Namely it depends on what you want to do with the generated clusters. Large maps produce a large number of small but "compact" clusters (records assigned to each cluster are quite similar). Small maps produce less but more generilized clusters. A "right number of clusters" doesn't exists, especially in real world datasets. It all depends on the detail which you want to examine your dataset.
I have written a function that, with the data set as input, returns the grid size. I rewrote it from the som_topol_struct() function of Matlab's Self Organizing Maps Toolbox into a R function.
topology=function(data)
{
#Determina, para lattice hexagonal, el número de neuronas (munits) y su disposición (msize)
D=data
# munits: número de hexágonos
# dlen: número de sujetos
dlen=dim(data)[1]
dim=dim(data)[2]
munits=ceiling(5*dlen^0.5) # Formula Heurística matlab
#munits=100
#size=c(round(sqrt(munits)),round(munits/(round(sqrt(munits)))))
A=matrix(Inf,nrow=dim,ncol=dim)
for (i in 1:dim)
{
D[,i]=D[,i]-mean(D[is.finite(D[,i]),i])
}
for (i in 1:dim){
for (j in i:dim){
c=D[,i]*D[,j]
c=c[is.finite(c)];
A[i,j]=sum(c)/length(c)
A[j,i]=A[i,j]
}
}
VS=eigen(A)
eigval=sort(VS$values)
if (eigval[length(eigval)]==0 | eigval[length(eigval)-1]*munits<eigval[length(eigval)]){
ratio=1
}else{
ratio=sqrt(eigval[length(eigval)]/eigval[length(eigval)-1])}
size1=min(munits,round(sqrt(munits/ratio*sqrt(0.75))))
size2=round(munits/size1)
return(list(munits=munits,msize=sort(c(size1,size2),decreasing=TRUE)))
}
hope it helps...
Iván Vallés-Pérez
I don't have a reference for it, but I would suggest starting off by using approximately 10 SOM neurons per expected class in your dataset. For example, if you think your dataset consists of 8 separate components, go for a map with 9x9 neurons. This is completely just a ballpark heuristic though.
If you'd like the data to drive the topology of your SOM a bit more directly, try one of the SOM variants that change topology during training:
Growing SOM
Growing Neural Gas
Unfortunately these algorithms involve even more parameter tuning than plain SOM, but they might work for your application.
Kohenon has written on the issue of selecting parameters and map size for SOM in his book "MATLAB Implementations and Applications of the Self-Organizing Map". In some cases, he suggest the initial values can be arrived at after testing several sizes of the SOM to check that the cluster structures were shown with sufficient resolution and statistical accuracy.
my suggestion would be the following
SOM is distantly related to correspondence analysis. In statistics, they use 5*r^2 as a rule of thumb, where r is the number of rows/columns in a square setup
usually, one should use some criterion that is based on the data itself, meaning that you need some criterion for estimating the homogeneity. If a certain threshold would be violated, you would need more nodes. For checking the homogeneity you would need some records per node. Agai, from statistics you could learn that for simple tests (small number of variables) you would need around 20 records, for more advanced tests on some variables at least 8 records.
remember that the SOM represents a predictive model. So validation is the key, absolutely mandatory. Yet, validation of predictive models (see typeI / II error entry in Wiki) is a subject on its own. And the acceptable risk as well as the risk structure also depend fully on your purpose.
You may test the dynamics of the error rate of the model by reducing its size more and more. Then take the smallest one with acceptable error.
It is a strength of the SOM to allow for empty nodes. Yet, there should not be too much of them. Let me say, less than 5%.
Taken all together, from experience, I would recommend the following criterion a minimum of the absolute number of 8..10 records, but those should not be more than 5% of all clusters.
Those 5% rule is of of course a heuristics, which however can be justified by the general usage of the confidence level in statistical tests. You may choose any percentage from 1% to 5%.
I am using Support Vector Machines for document classification. My feature set for each document is a tf-idf vector. I have M documents with each tf-idf vector of size N.
Giving M * N matrix.
The size of M is just 10 documents and tf-idf vector is 1000 word vector. So my features are much larger than number of documents. Also each word occurs in either 2 or 3 documents. When i am normalizing each feature ( word ) i.e. column normalization in [0,1] with
val_feature_j_row_i = ( val_feature_j_row_i - min_feature_j ) / ( max_feature_j - min_feature_j)
It either gives me 0, 1 of course.
And it gives me bad results. I am using libsvm, with rbf function C = 0.0312, gamma = 0.007815
Any recommendations ?
Should i include more documents ? or other functions like sigmoid or better normalization methods ?
The list of things to consider and correct is quite long, so first of all I would recommend some machine-learning reading before trying to face the problem itself. There are dozens of great books (like ie. Haykin's "Neural Networks and Learning Machines") as well as online courses, which will help you with such basics, like those listed here: http://www.class-central.com/search?q=machine+learning .
Getting back to the problem itself:
10 documents is rows of magnitude to small to get any significant results and/or insights into the problem,
there is no universal method of data preprocessing, you have to analyze it through numerous tests and data analytics,
SVMs are parametrical models, you cannot use a single C and gamma values and expect any reasonable results. You have to check dozens of them to even get a clue "where to search". The most simple method for doing so is so called grid search,
1000 of features is a great number of dimensions, this suggest that using a kernel, which implies infinitely dimensional feature space is quite... redundant - it would be a better idea to first analyze simplier ones, which have smaller chance to overfit (linear or low degree polynomial)
finally is tf*idf a good choice if "each word occurs in 2 or 3 documents"? It can be doubtfull, unless what you actually mean is 20-30% of documents
finally why is simple features squashing
It either gives me 0, 1 of course.
it should result in values in [0,1] interval, not just its limits. So if this is a case you are probably having some error in your implementation.
I am a bit confused about the namings in the SVM. I am using this library LibSVM. There are so many parameters that can be set. Does anyone know which of these is the slack variable?
thx
The "slack variable" is C in c-svm and nu in nu-SVM. These both serve the same function in their respective formulations - controlling the tradeoff between a wide margin and classifier error. In the case of C, one generally test it in orders of magnitude, say 10^-4, 10^-3, 10^-2,... to 1, 5 or so. nu is a number between 0 and 1, generally from .1 to .8, which controls the ratio of support vectors to data points. When nu is .1, the margin is small, the number of support vectors will be a small percentage of the number of data points. When nu is .8, the margin is very large and most of the points will fall in the margin.
The other things to consider are your choice of kernel (linear, RBF, sigmoid, polynomial) and the parameters for the chosen kernel. Generally one has to do a lot of experimenting to find the best combination of parameters. However, be careful of over-fitting to your dataset.
Burges wrote a great tutorial: A Tutorial on Support Vector Machines for Pattern
Recognition
But if you mostly just want to know how to USE it and less about how it works, read "A Practical Guide to Support Vector Classication" by Chih-Wei Hsu, Chih-Chung Chang, and Chih-Jen Lin (authors of libsvm)
First decide which type of SVM are u intending to use: C-SVC, nu-SVC , epsilon-SVR or nu-SVR. In my opinion u need to vary C and gamma most of the time... the rest are usually fixed..