Let's assume we have three columns, with binary features (0,1). One row in dataset is 0,0,0 with label 0.
The problem I am facing is:
When assigning weights to this row and activating sigmoid function, I'll always receive 0.5, because an array of zeros products on any weight with 0.
How to overcome this issue?
In addition to multiplication by a weight matrix, you can also add a bias (which is how it is typically done in neural networks), and hence you won't necessarily get a zero vector. You could also add more hidden layers (but as I said, even adding a bias vector will resolve the issue you mentioned).
Related
Suppose there is a matrix B, where its size is a 500*1000 double(Here, 500 represents the number of observations and 1000 represents the number of features).
sigma is the covariance matrix of B, and D is a diagonal matrix whose diagonal elements are the eigenvalues of sigma. Assume A is the eigenvectors of the covariance matrix sigma.
I have the following questions:
I need to select the first k = 800 eigenvectors corresponding to the eigenvalues with the largest magnitude to rank the selected features. The final matrix named Aq. How can I do this in MATLAB?
What is the meaning of these selected eigenvectors?
It seems the size of the final matrix Aq is 1000*800 double once I calculate Aq. The time points/observation information of 500 has disappeared. For the final matrix Aq, what does the value 1000 in matrix Aq represent now? Also, what does the value 800 in matrix Aq represent now?
I'm assuming you determined the eigenvectors from the eig function. What I would recommend to you in the future is to use the eigs function. This not only computes the eigenvalues and eigenvectors for you, but it will compute the k largest eigenvalues with their associated eigenvectors for you. This may save computational overhead where you don't have to compute all of the eigenvalues and associated eigenvectors of your matrix as you only want a subset. You simply supply the covariance matrix of your data to eigs and it returns the k largest eigenvalues and eigenvectors for you.
Now, back to your problem, what you are describing is ultimately Principal Component Analysis. The mechanics behind this would be to compute the covariance matrix of your data and find the eigenvalues and eigenvectors of the computed result. It has been known that doing it this way is not recommended due to numerical instability with computing the eigenvalues and eigenvectors for large matrices. The most canonical way to do this now is via Singular Value Decomposition. Concretely, the columns of the V matrix give you the eigenvectors of the covariance matrix, or the principal components, and the associated eigenvalues are the square root of the singular values produced in the diagonals of the matrix S.
See this informative post on Cross Validated as to why this is preferred:
https://stats.stackexchange.com/questions/79043/why-pca-of-data-by-means-of-svd-of-the-data
I'll throw in another link as well that talks about the theory behind why the Singular Value Decomposition is used in Principal Component Analysis:
https://stats.stackexchange.com/questions/134282/relationship-between-svd-and-pca-how-to-use-svd-to-perform-pca
Now let's answer your question one at a time.
Question #1
MATLAB generates the eigenvalues and the corresponding ordering of the eigenvectors in such a way where they are unsorted. If you wish to select out the largest k eigenvalues and associated eigenvectors given the output of eig (800 in your example), you'll need to sort the eigenvalues in descending order, then rearrange the columns of the eigenvector matrix produced from eig then select out the first k values.
I should also note that using eigs will not guarantee sorted order, so you will have to explicitly sort these too when it comes down to it.
In MATLAB, doing what we described above would look something like this:
sigma = cov(B);
[A,D] = eig(sigma);
vals = diag(D);
[~,ind] = sort(abs(vals), 'descend');
Asort = A(:,ind);
It's a good thing to note that you do the sorting on the absolute value of the eigenvalues because scaled eigenvalues are also eigenvalues themselves. These scales also include negatives. This means that if we had a component whose eigenvalue was, say -10000, this is a very good indication that this component has some significant meaning to your data, and if we sorted purely on the numbers themselves, this gets placed near the lower ranks.
The first line of code finds the covariance matrix of B, even though you said it's already stored in sigma, but let's make this reproducible. Next, we find the eigenvalues of your covariance matrix and the associated eigenvectors. Take note that each column of the eigenvector matrix A represents one eigenvector. Specifically, the ith column / eigenvector of A corresponds to the ith eigenvalue seen in D.
However, the eigenvalues are in a diagonal matrix, so we extract out the diagonals with the diag command, sort them and figure out their ordering, then rearrange A to respect this ordering. I use the second output of sort because it tells you the position of where each value in the unsorted result would appear in the sorted result. This is the ordering we need to rearrange the columns of the eigenvector matrix A. It's imperative that you choose 'descend' as the flag so that the largest eigenvalue and associated eigenvector appear first, just like we talked about before.
You can then pluck out the first k largest vectors and values via:
k = 800;
Aq = Asort(:,1:k);
Question #2
It's a well known fact that the eigenvectors of the covariance matrix are equal to the principal components. Concretely, the first principal component (i.e. the largest eigenvector and associated largest eigenvalue) gives you the direction of the maximum variability in your data. Each principal component after that gives you variability of a decreasing nature. It's also good to note that each principal component is orthogonal to each other.
Here's a good example from Wikipedia for two dimensional data:
I pulled the above image from the Wikipedia article on Principal Component Analysis, which I linked you to above. This is a scatter plot of samples that are distributed according to a bivariate Gaussian distribution centred at (1,3) with a standard deviation of 3 in roughly the (0.878, 0.478) direction and of 1 in the orthogonal direction. The component with a standard deviation of 3 is the first principal component while the one that is orthogonal is the second component. The vectors shown are the eigenvectors of the covariance matrix scaled by the square root of the corresponding eigenvalue, and shifted so their tails are at the mean.
Now let's get back to your question. The reason why we take a look at the k largest eigenvalues is a way of performing dimensionality reduction. Essentially, you would be performing a data compression where you would take your higher dimensional data and project them onto a lower dimensional space. The more principal components you include in your projection, the more it will resemble the original data. It actually begins to taper off at a certain point, but the first few principal components allow you to faithfully reconstruct your data for the most part.
A great visual example of performing PCA (or SVD rather) and data reconstruction is found by this great Quora post I stumbled upon in the past.
http://qr.ae/RAEU8a
Question #3
You would use this matrix to reproject your higher dimensional data onto a lower dimensional space. The number of rows being 1000 is still there, which means that there were originally 1000 features in your dataset. The 800 is what the reduced dimensionality of your data would be. Consider this matrix as a transformation from the original dimensionality of a feature (1000) down to its reduced dimensionality (800).
You would then use this matrix in conjunction with reconstructing what the original data was. Concretely, this would give you an approximation of what the original data looked like with the least amount of error. In this case, you don't need to use all of the principal components (i.e. just the k largest vectors) and you can create an approximation of your data with less information than what you had before.
How you reconstruct your data is very simple. Let's talk about the forward and reverse operations first with the full data. The forward operation is to take your original data and reproject it but instead of the lower dimensionality, we will use all of the components. You first need to have your original data but mean subtracted:
Bm = bsxfun(#minus, B, mean(B,1));
Bm will produce a matrix where each feature of every sample is mean subtracted. bsxfun allows the subtraction of two matrices in unequal dimension provided that you can broadcast the dimensions so that they can both match up. Specifically, what will happen in this case is that the mean of each column / feature of B will be computed and a temporary replicated matrix will be produced that is as large as B. When you subtract your original data with this replicated matrix, the effect will subtract every data point with their respective feature means, thus decentralizing your data so that the mean of each feature is 0.
Once you do this, the operation to project is simply:
Bproject = Bm*Asort;
The above operation is quite simple. What you are doing is expressing each sample's feature as a linear combination of principal components. For example, given the first sample or first row of the decentralized data, the first sample's feature in the projected domain is a dot product of the row vector that pertains to the entire sample and the first principal component which is a column vector.. The first sample's second feature in the projected domain is a weighted sum of the entire sample and the second component. You would repeat this for all samples and all principal components. In effect, you are reprojecting the data so that it is with respect to the principal components - which are orthogonal basis vectors that transform your data from one representation to another.
A better description of what I just talked about can be found here. Look at Amro's answer:
Matlab Principal Component Analysis (eigenvalues order)
Now to go backwards, you simply do the inverse operation, but a special property with the eigenvector matrix is that if you transpose this, you get the inverse. To get the original data back, you undo the operation above and add the means back to the problem:
out = bsxfun(#plus, Bproject*Asort.', mean(B, 1));
You want to get the original data back, so you're solving for Bm with respect to the previous operation that I did. However, the inverse of Asort is just the transpose here. What's happening after you perform this operation is that you are getting the original data back, but the data is still decentralized. To get the original data back, you must add the means of each feature back into the data matrix to get the final result. That's why we're using another bsxfun call here so that you can do this for each sample's feature values.
You should be able to go back and forth from the original domain and projected domain with the above two lines of code. Now where the dimensionality reduction (or the approximation of the original data) comes into play is the reverse operation. What you need to do first is project the data onto the bases of the principal components (i.e. the forward operation), but now to go back to the original domain where we are trying to reconstruct the data with a reduced number of principal components, you simply replace Asort in the above code with Aq and also reduce the amount of features you're using in Bproject. Concretely:
out = bsxfun(#plus, Bproject(:,1:k)*Aq.', mean(B, 1));
Doing Bproject(:,1:k) selects out the k features in the projected domain of your data, corresponding to the k largest eigenvectors. Interestingly, if you just want the representation of the data with regards to a reduced dimensionality, you can just use Bproject(:,1:k) and that'll be enough. However, if you want to go forward and compute an approximation of the original data, we need to compute the reverse step. The above code is simply what we had before with the full dimensionality of your data, but we use Aq as well as selecting out the k features in Bproject. This will give you the original data that is represented by the k largest eigenvectors / eigenvalues in your matrix.
If you'd like to see an awesome example, I'll mimic the Quora post that I linked to you but using another image. Consider doing this with a grayscale image where each row is a "sample" and each column is a feature. Let's take the cameraman image that's part of the image processing toolbox:
im = imread('camerman.tif');
imshow(im); %// Using the image processing toolbox
We get this image:
This is a 256 x 256 image, which means that we have 256 data points and each point has 256 features. What I'm going to do is convert the image to double for precision in computing the covariance matrix. Now what I'm going to do is repeat the above code, but incrementally increasing k at each go from 3, 11, 15, 25, 45, 65 and 125. Therefore, for each k, we are introducing more principal components and we should slowly start to get a reconstruction of our data.
Here's some runnable code that illustrates my point:
%%%%%%%// Pre-processing stage
clear all;
close all;
%// Read in image - make sure we cast to double
B = double(imread('cameraman.tif'));
%// Calculate covariance matrix
sigma = cov(B);
%// Find eigenvalues and eigenvectors of the covariance matrix
[A,D] = eig(sigma);
vals = diag(D);
%// Sort their eigenvalues
[~,ind] = sort(abs(vals), 'descend');
%// Rearrange eigenvectors
Asort = A(:,ind);
%// Find mean subtracted data
Bm = bsxfun(#minus, B, mean(B,1));
%// Reproject data onto principal components
Bproject = Bm*Asort;
%%%%%%%// Begin reconstruction logic
figure;
counter = 1;
for k = [3 11 15 25 45 65 125 155]
%// Extract out highest k eigenvectors
Aq = Asort(:,1:k);
%// Project back onto original domain
out = bsxfun(#plus, Bproject(:,1:k)*Aq.', mean(B, 1));
%// Place projection onto right slot and show the image
subplot(4, 2, counter);
counter = counter + 1;
imshow(out,[]);
title(['k = ' num2str(k)]);
end
As you can see, the majority of the code is the same from what we have seen. What's different is that I loop over all values of k, project back onto the original space (i.e. computing the approximation) with the k highest eigenvectors, then show the image.
We get this nice figure:
As you can see, starting with k=3 doesn't really do us any favours... we can see some general structure, but it wouldn't hurt to add more in. As we start increasing the number of components, we start to get a clearer picture of what the original data looks like. At k=25, we actually can see what the cameraman looks like perfectly, and we don't need components 26 and beyond to see what's happening. This is what I was talking about with regards to data compression where you don't need to work on all of the principal components to get a clear picture of what's going on.
I'd like to end this note by referring you to Chris Taylor's wonderful exposition on the topic of Principal Components Analysis, with code, graphs and a great explanation to boot! This is where I got started on PCA, but the Quora post is what solidified my knowledge.
Matlab - PCA analysis and reconstruction of multi dimensional data
I have a matrix X, the size of which is 100*2000 double. I want to know which kind of scaling technique is applied to matrix X in the following command, and why it does not use z-score to do scaling?
X = X./repmat(sqrt(sum(X.^2)),size(X,1),1);
That scaling comes from linear algebra. That's what we call normalizing by producing a unit vector. Assuming that each row is an observation and each column is a feature, what's happening here is that we are going through every observation that you collected and normalizing each feature value over all observations such that the overall length / magnitude of a particular feature for all observations is set to 1.
The bottom division takes a look at each feature and determines the norm or magnitude of the feature over all observations. Once you find these magnitudes, you then take each feature for each observation and divide by their respective magnitudes.
The reason why unit vectors are often employed is to describe a point in feature space with respect to a set of basis vectors. Normalizing by producing unit vectors gives you the smallest possible way to represent one component in feature space and so what's probably happening here is that the observations are now being transformed such that each component / feature is being represented in terms of a set of basis vectors. Each basis vector is one feature in the data.
Check out the Wikipedia article on Unit Vectors for more details: http://en.wikipedia.org/wiki/Unit_vector
According to libsvm faqs, the following one-line code scale each feature to the range of [0,1] in Matlab
(data - repmat(min(data,[],1),size(data,1),1))*spdiags(1./(max(data,[],1)-min(data,[],1))',0,size(data,2),size(data,2))
so I'm using this code:
v_feature_trainN=(v_feature_train - repmat(mini,size(v_feature_train,1),1))*spdiags(1./(maxi-mini)',0,size(v_feature_train,2),size(v_feature_train,2));
v_feature_testN=(v_feature_test - repmat(mini,size(v_feature_test,1),1))*spdiags(1./(maxi-mini)',0,size(v_feature_test,2),size(v_feature_test,2));
where I use the first one to train the classifier and the second one to classify...
In my humble opinion scaling should be performed by:
i.e.:
v_feature_trainN2=(v_feature_train -min(v_feature_train(:)))./(max(v_feature_train(:))-min((v_feature_train(:))));
v_feature_test_N2=(v_feature_test -min(v_feature_train(:)))./(max(v_feature_train(:))-min((v_feature_train(:))));
Now I compared the classification results using these two scaling methods and the first one outperforms the second one.
The question are:
1) What exactly does the first method? I didn't understand it.
2) Why the code suggested by libsvm outperforms the second one (e.g. 80% vs 60%)?
Thank you so much in advance
First of all:
The code described in the libsvm does something different than your code:
It maps every column independently onto the interval [0,1].
Your code however uses the global min and max to map all the columns using the same affine transformation instead of a separate transformation for each column.
The first code works in the following way:
(data - repmat(min(data,[],1),size(data,1),1))
This subtracts each column's minimum from the entire column. It does this by computing the row vector of minima min(data,[],1) which is then replicated to build a matrix the same size as data. Then it is subtracted from data.
spdiags(1./(max(data,[],1)-min(data,[],1))',0,size(data,2),size(data,2))
This generates a diagonal matrix. The entry (i,i) of this matrix is 1 divided by the difference of the maximum and the minimum of the ith column: max(data(:,i))-min(data(:,i)).
The right multiplication of this diagonal matrix means: Multiply each column of the left matrix with the corresponding diagonal entry. This effectively divides column i by max(data(:,i))-min(data(:,i)).
Instead of using a sparse diagonal matrix, you could do this even more efficiently with bsxfun:
bsxfun(#rdivide, ...
bsxfun(#minus, ...
data, min(data,[],1)), ...
max(data,[],1)-min(data,[],1))
Which is the matlab way of writing:
Divide:
The difference of:
each column and its respective minimum
by the difference of each column's max and min.
I know this has already been answered correctly, but I would like to present another solution that I think is also correct and I found more intuitive/shorther then the one presented by knedlsepp. I am new to matlab and as I was studying knedlsepp solution, I found it more intuitive to solve this problem with the following formula:
function [ output ] = feature_scaling( y)
output = (y - repmat(min(y),size(y,1),1)) * diag(1./(max(y) - min(y)));
end
I find it a bit easier to use diag this way instead of spdiags, but I believe it produces the same result for the purpose of this excercise.
Multiplying the first term by the second, effectively divides each member of the matrix (Y-min(Y)) by the scalar value 1/(max(y)-min(y)), achieving the desired result.
In case someone prefers a shorter version, maybe this can be of help.
I have a set of samples, S, and I want to find its PDF. The problem is when I use ksdensity I get values greater than one!
[f,xi] = ksdensity(S)
In array f, most of the values are greater than one! Would you please tell me what the problem can be? Thanks for your help.
For example:
S=normrnd(0.3035, 0.0314,1,1000);
ksdensity(S)
ksdensity, as the name says, estimates a probability density function over a continuous variable. Probability densities can be larger than 1, they can actually have arbitrary values from zero upwards. The constraint on probabilities is that their sum over an exhaustive range of possibilities has to be 1. For probability densities, the constraint is that the integral over the whole range of values is 1.
A crude approximation of an integral of the pdf estimated by ksdensity can be obtained in Matlab like this:
sum(f) * min(diff(xi))
assuming that the values in xi are equally spaced. The value of this expression should be approximately 1.
If in your application you believe this approximation is not close enough to 1, you might want to specify the grid of estimation points (second parameter pts) such that the spacing is finer or the range is wider than the one automatically generated by ksdensity.
I have been using Matlab's normxcorr2 function to do template matching with images by performing normalized cross correlation. To find the maximum correspondence between a template and an image, one can simply run normxcorr2 and then find the maximum absolute value of all the values returned by normxcorr2 (the function returns values between -1.0 and 1.0).
From a quick Google search, I found out that a negative correlation coefficient means an inverse relationship between two variables (e.g. as x increases, y decreases), and that a positive correlation coefficient means the opposite (e.g. as x increases, y increases). How does this apply to image template matching? That is, what does a negative value from normxcorr2 mean conceptually with respect to template matching?
View normalized cross correlation as a normalized vector dot product. If the angle between two vectors is zero, their dot product will be 1; if they are facing in the opposite direction, then their dot product with be negative 1. This is idea is actually direct if you take the array and stack the column end to end. The result is essentially a dot product between two vectors.
Also just as a personal anecdote: what confused me about template matching at first, was intuitively I believed element wise subtraction of two images would be a good metric for image similarity. When I first saw cross correlation, I wondered why it used element wise multiplication. Then I realized that the later operation is the same thing as a vector dot product. The vector dot product, as I mentioned before, indicates when two vectors are pointing in the same direction. In your case, the two vectors are normalized first; hence why the range is from -1 to 1. If you want to read more about the implementation, "Fast Normalized Cross Correlation" by J.P. Lewis is a classical paper on the subject.
Check the formula
on wikipedia.
When f(x, y) - mean(f) and t(x,y) - mean(t) have different sign the result of an addendum will be negative (std is always positive). If there are a lot of such (x,y) then the whole sum will also be negative. You may think that if 1.0 means that one image is equal to another. -1.0 means that one image is a negative of another (try to find normxcorr2(x, -x))