How to 'smooth' data and calculate line gradient? - algorithm

I'm reading data from a device which measures distance. My sample rate is high so that I can measure large changes in distance (i.e. velocity) but this means that, when the velocity is low, the device delivers a number of measurements which are identical (due to the granularity of the device). This results in a 'stepped' curve.
What I need to do is to smooth the curve in order to calculate the velocity. Following that I then need to calculate the acceleration.
How to best go about this?
(Sample rate up to 1000Hz, calculation rate of 10Hz would be ok. Using C# in VS2005)

The wikipedia entry from moogs is a good starting point for smoothing the data. But it does not help you in making a decision.
It all depends on your data, and the needed processing speed.
Moving Average
Will flatten the top values. If you are interrested in the minimum and maximum value, don't use this. Also I think using the moving average will influence your measurement of the acceleration, since it will flatten your data (a bit), thereby acceleration will appear to be smaller. It all comes down to the needed accuracy.
Savitzky–Golay
Fast algorithm. As fast as the moving average. That will preserve the heights of peaks. Somewhat harder to implement. And you need the correct coefficients. I would pick this one.
Kalman filters
If you know the distribution, this can give you good results (it is used in GPS navigation systems). Maybe somewhat harder to implement. I mention this because I have used them in the past. But they are probably not a good choice for a starter in this kind of stuff.
The above will reduce noise on your signal.
Next you have to do is detect the start and end point of the "acceleration". You could do this by creating a Derivative of the original signal. The point(s) where the derivative crosses the Y-axis (zero) are probably the peaks in your signal, and might indicate the start and end of the acceleration.
You can then create a second degree derivative to get the minium and maximum acceleration itself.

You need a smoothing filter, the simplest would be a "moving average": just calculate the average of the last n points.
The question here is, how to determine n, can you tell us more about your application?
(There are other, more complicated filters. They vary on how they preserve the input data. A good list is in Wikipedia)
Edit!: For 10Hz, average the last 100 values.

Moving averages are generally terrible - but work well for white noise. Both moving averages & Savitzky-Golay both boil down to a correlation - and therefore are very fast and could be implemented in real time. If you need higher order information like first and second derivatives - SG is a good right choice. The magic of SG lies in the constant correlation coefficients needed for the filter - once you have decided the length and degree of polynomial to fit locally, the coefficients need only to be found once. You can compute them using R (sgolay) or Matlab.
You can also estimate a noisy signal's first derivative via the Savitzky-Golay best-fit polynomials - these are sometimes called Savitzky-Golay derivatives - and typically give a good estimate of the first derivative.
Kalman filtering can be very effective, but it's heavier computationally - it's hard to beat a short convolution for speed!
Paul
CenterSpace Software

In addition to the above articles, have a look at Catmull-Rom Splines.

You could use a moving average to smooth out the data.

In addition to GvSs excellent answer above you could also consider smoothing / reducing the stepping effect of your averaged results using some general curve fitting such as cubic or quadratic splines.

Related

SGM Disparity subpixel estimation - how to?

Some weeks ago I've implemented a simple block matching stereo algorithm but the results had been bad. So I've searched on the Internet to find better algorithms. There I found the semi global matching (SGM), published by Heiko Hirschmueller. It gets one of the best results in relation to its processing time.
I've implemented the algorithm and got really good results (compared to simple block matching) as you can see here:
I've reprojected the 2D points to 3D by using the calculated disparity values with the following result
At the end of SGM I have an array with aggregated costs for each pixel. The disparity is equivalent to the index with the lowest cost value.
The problem is, that searching for the minimum only returns discrete values. This results in individually layers in the point-cloud. In other words: Round surfaces are cut into many layers (see point cloud).
Heiko mentioned in his paper, that it would be easy to get sub-pixel accuracy by fitting a polynomial function into the cost array and take the lowest point as disparity.
The problem is not bound to stereo vision, so in other words the task is the following:
given: An array of values, representing a polynomial function.
wanted: The lowest point of the polynomial function.
I don't have any idea how to do this. I need a fast algorithm, because I have to run this code for every pixel in the Image
For example: 500x500 Pixel with 60-200 costs each => Algorithm has to run 15000000-50000000 times!!).
I don't need a real time solution! My current SGM implementation (L2R and R2L matching, no cuda or multi-threading yet) takes about 20 seconds to process an image with 500x500 pixels ;).
I don't ask for libraries! I try to implement my own independent computer vision library :).
Thank you for your help!
With kind regards,
Andreas
Finding the exact lowest point in a general polynomial is a hard problem, since it is equivalent to finding the root of the derivative of the polynomial. In particular, if your polynomial is of degree 6, the derivative is a quintic polynomial, which is known not to be solvable by radical. You therefore need to either: fit the function using restricted families for which computing the roots of the derivatives e.g. the integrals of prod_i(x-ri)p(q) where deg(p)<=4, OR
using an iterative method to find an APPROXIMATE minimum, (newton's method, gradient descent).

how to calculate the predicted motion vector in ARPS

I'm doing a video homework in which you program one of the famous video compression algorithms, I chose ARPS(adaptive rood pattern search).
Now if I understand it right I must first divide the image into macroblocks, I've already done that, second calculate the pmv(predicted motion vector) by taking the motion vector of the left neighboring macroblock(type D, there are other types in which you take the above or left-above etc, according to some paper they don't differ much in quality).
Last use pmv to calculate the mv of the current macroblock.
If I understand it correctly I have to calculate the first column of macroblocks using other algorithms(NTSS or FSS or etc) and then use that column to calculate the rest.
What will happen if my first column didn't move ? pmv=(0,0) and applying the algorithm as I understand it from wikipedia results in all mvs being (0,0) (aka first column didn't change=nothing changed !!!)
I doubt I understand the algorithm correctly and for some reason many papers don't address those issues, so can you shed some light on it ? I can implement it very well after that.
PS
this is a university homework and I'm at software-engineering department (not AI department) so no AI algorithms please .
As I know strength of ARPS algorithm is adaptive rood size, which base on predicted motion vector (mv). Rood is used as the first step of this method, and may dramatical reduce over estimation by jumping in the right place. Second step is to fine the best point(s) by small diamond or other fixed patterns you like.
So if you predict zero mv, you simple apply fine estimation (second step).
In practice, zero mv mean small metric value of estimated block (by SAD or other metric). Natural static pictures always have some deviation of adjacent (in temporal plane) samples and any metric produce some value. It is decision of you implementation marking this block by zero moving or with small motion vector.
About the column in you implementation. For first block in the non first row you may use mv of above block. that reduce you total calculations.
In any case, you may check is you changes of implementation good or bad, by apply motion compensation (inverse of motion estimation) and calculate with original picture by calculating the PSNR metric.

strategy for fitting angle data

I have a set of angles. The distribution could be roughly described as:
there a are usually several values very close (0.0-1.0 degree apart) to the correct solution
there are also noisy values being very far from the correct result, even opposite direction
is there a common solution/strategy for such a problem?
For multidimensional data, I would use RANSAC - but I have the impression that it is unusual to apply Ransac on 1-dimensional data. Another problem is computing the mean of an angle. I read some other posts about how to calculate the mean of angles by using vectors, but I just wonder if there isn't a particular fitting solution which deals with both issues already.
You can use RANSAC even in this case, all the necessary conditions (minimal samples, error of a data point, consensus set) are met. Your minimal sample will be 1 point, a randomly picked angle (although you can try all of them, might be fast enough). Then, all the angles (data points) with error (you can use just absolute distance, modulo 360) less than some threshold (e.g. 1 degree), will be considered as inliers, i.e. within the consensus set.
If you want to play with it a bit more, you can make the results more stable by adding some local optimisation, see e.g.:
Lebeda, Matas, Chum: Fixing the Locally Optimized RANSAC, BMVC 2012.
You could try another approaches, e.g. median, or fitting a mixture of Gaussian and uniform distribution, but you would have to deal with the periodicity of the signal somehow, so I guess RANSAC should be your choice.

Pseudo code for Blockwise Non Local Means noise reduction algorithm

I have implemented a nice algorithm ("Non Local Means") for reducing noise in image.
It is based on it's Matlab implementation.
The problem with NLMeans is that the original algorithm is slow even on compiled languages like c/c++ and i am trying to run it using scripting language.
One of best solutions is to use improved Blockwise NLMeans algorithm which is ~60-80 times faster. The problem is that the paper which describes it is written in a complex mathematical language and it's really hard for me to understand an idea and program it
(yes, i didn't learn math at college).
That is why i am desperately looking for a pseudo code of this algorithm.
(modification of original Matlab implementation would be just perfect)
I admit, I was intrigued until I saw the result – 260+ seconds on a dual core, and that doesn't assume the overhead of a scripting language, and that's for the Optimized Block Non Local Means filter.
Let me break down the math for you – My idea of pseudo-code is writing in Ruby.
Non Local Means Filtering
Assume an image that's 200 x 100 pixels (20000 pixels total), which is a pretty tiny image. We're going to have to go through 20,000 pixels and evaluate each one on the weighted average of the other 19,999 pixels: [Sorry about the spacing, but the equation is interpreted as a link without it]
NL [v] (i) = ∑ w(i,j)v(j) [j ∈ I]
where 0 ≤ w(i,j) ≤ 1 and ∑j w(i,j) = 1
Understandably, this last part can be a little confusing, but this is really nothing more than a convolution filter the size of the whole image being applied to each pixel.
Blockwise Non Local Means Filtering
The blockwise implementation takes overlapping sets of voxels (volumetric pixels - the implementation you pointed us to is for 3D space). Presumably, taking a similar approach, you could apply this to 2D space, taking sets of overlapping pixels. Let's see if we can describe this...
NL [v] (ijk) = 1/|Ai|∑ w(ijk, i)v(i)
Where A is a vector of the pixels to be estimated, and similar circumstances as above are applied.
[NB: I may be slightly off; It's been a few years since I did heavy image processing]
Algorithm
In all likelihood, we're talking about reducing complexity of the algorithm at a minimal cost to reduction quality. The larger the sample vector, the higher the quality as well as the higher the complexity. By overlapping then averaging the sample vectors from the image then applying that weighted average to each pixel we're looping through the image far fewer times.
Loop through the image to collect the sample vectors and store their weighted average to an array.
Apply each weighted average (a number between 0 and 1) to each pixel times the pixels value.
Pretty simple, but the processing time is going to be horrid with larger images.
Final Thoughts
You're going to have to make some tough decisions. If you're going to use a scripting language, you're already dealing with significant interpretive overhead. It's far from optimal to use a scripting language for heavy duty image processing. If you're not processing medical images, in all likelihood, there are far better algorithms to use with lesser O's.
Hope this is helpful... I'm not terribly good at making a point clearly and concisely, so if I can clarify anything, let me know.

Nearest neighbors in high-dimensional data?

I have asked a question a few days back on how to find the nearest neighbors for a given vector. My vector is now 21 dimensions and before I proceed further, because I am not from the domain of Machine Learning nor Math, I am beginning to ask myself some fundamental questions:
Is Euclidean distance a good metric for finding the nearest neighbors in the first place? If not, what are my options?
In addition, how does one go about deciding the right threshold for determining the k-neighbors? Is there some analysis that can be done to figure this value out?
Previously, I was suggested to use kd-Trees but the Wikipedia page clearly says that for high-dimensions, kd-Tree is almost equivalent to a brute-force search. In that case, what is the best way to find nearest-neighbors in a million point dataset efficiently?
Can someone please clarify the some (or all) of the above questions?
I currently study such problems -- classification, nearest neighbor searching -- for music information retrieval.
You may be interested in Approximate Nearest Neighbor (ANN) algorithms. The idea is that you allow the algorithm to return sufficiently near neighbors (perhaps not the nearest neighbor); in doing so, you reduce complexity. You mentioned the kd-tree; that is one example. But as you said, kd-tree works poorly in high dimensions. In fact, all current indexing techniques (based on space partitioning) degrade to linear search for sufficiently high dimensions [1][2][3].
Among ANN algorithms proposed recently, perhaps the most popular is Locality-Sensitive Hashing (LSH), which maps a set of points in a high-dimensional space into a set of bins, i.e., a hash table [1][3]. But unlike traditional hashes, a locality-sensitive hash places nearby points into the same bin.
LSH has some huge advantages. First, it is simple. You just compute the hash for all points in your database, then make a hash table from them. To query, just compute the hash of the query point, then retrieve all points in the same bin from the hash table.
Second, there is a rigorous theory that supports its performance. It can be shown that the query time is sublinear in the size of the database, i.e., faster than linear search. How much faster depends upon how much approximation we can tolerate.
Finally, LSH is compatible with any Lp norm for 0 < p <= 2. Therefore, to answer your first question, you can use LSH with the Euclidean distance metric, or you can use it with the Manhattan (L1) distance metric. There are also variants for Hamming distance and cosine similarity.
A decent overview was written by Malcolm Slaney and Michael Casey for IEEE Signal Processing Magazine in 2008 [4].
LSH has been applied seemingly everywhere. You may want to give it a try.
[1] Datar, Indyk, Immorlica, Mirrokni, "Locality-Sensitive Hashing Scheme Based on p-Stable Distributions," 2004.
[2] Weber, Schek, Blott, "A quantitative analysis and performance study for similarity-search methods in high-dimensional spaces," 1998.
[3] Gionis, Indyk, Motwani, "Similarity search in high dimensions via hashing," 1999.
[4] Slaney, Casey, "Locality-sensitive hashing for finding nearest neighbors", 2008.
I. The Distance Metric
First, the number of features (columns) in a data set is not a factor in selecting a distance metric for use in kNN. There are quite a few published studies directed to precisely this question, and the usual bases for comparison are:
the underlying statistical
distribution of your data;
the relationship among the features
that comprise your data (are they
independent--i.e., what does the
covariance matrix look like); and
the coordinate space from which your
data was obtained.
If you have no prior knowledge of the distribution(s) from which your data was sampled, at least one (well documented and thorough) study concludes that Euclidean distance is the best choice.
YEuclidean metric used in mega-scale Web Recommendation Engines as well as in current academic research. Distances calculated by Euclidean have intuitive meaning and the computation scales--i.e., Euclidean distance is calculated the same way, whether the two points are in two dimension or in twenty-two dimension space.
It has only failed for me a few times, each of those cases Euclidean distance failed because the underlying (cartesian) coordinate system was a poor choice. And you'll usually recognize this because for instance path lengths (distances) are no longer additive--e.g., when the metric space is a chessboard, Manhattan distance is better than Euclidean, likewise when the metric space is Earth and your distances are trans-continental flights, a distance metric suitable for a polar coordinate system is a good idea (e.g., London to Vienna is is 2.5 hours, Vienna to St. Petersburg is another 3 hrs, more or less in the same direction, yet London to St. Petersburg isn't 5.5 hours, instead, is a little over 3 hrs.)
But apart from those cases in which your data belongs in a non-cartesian coordinate system, the choice of distance metric is usually not material. (See this blog post from a CS student, comparing several distance metrics by examining their effect on kNN classifier--chi square give the best results, but the differences are not large; A more comprehensive study is in the academic paper, Comparative Study of Distance Functions for Nearest Neighbors--Mahalanobis (essentially Euclidean normalized by to account for dimension covariance) was the best in this study.
One important proviso: for distance metric calculations to be meaningful, you must re-scale your data--rarely is it possible to build a kNN model to generate accurate predictions without doing this. For instance, if you are building a kNN model to predict athletic performance, and your expectation variables are height (cm), weight (kg), bodyfat (%), and resting pulse (beats per minute), then a typical data point might look something like this: [ 180.4, 66.1, 11.3, 71 ]. Clearly the distance calculation will be dominated by height, while the contribution by bodyfat % will be almost negligible. Put another way, if instead, the data were reported differently, so that bodyweight was in grams rather than kilograms, then the original value of 86.1, would be 86,100, which would have a large effect on your results, which is exactly what you don't want. Probably the most common scaling technique is subtracting the mean and dividing by the standard deviation (mean and sd refer calculated separately for each column, or feature in that data set; X refers to an individual entry/cell within a data row):
X_new = (X_old - mu) / sigma
II. The Data Structure
If you are concerned about performance of the kd-tree structure, A Voronoi Tessellation is a conceptually simple container but that will drastically improve performance and scales better than kd-Trees.
This is not the most common way to persist kNN training data, though the application of VT for this purpose, as well as the consequent performance advantages, are well-documented (see e.g. this Microsoft Research report). The practical significance of this is that, provided you are using a 'mainstream' language (e.g., in the TIOBE Index) then you ought to find a library to perform VT. I know in Python and R, there are multiple options for each language (e.g., the voronoi package for R available on CRAN)
Using a VT for kNN works like this::
From your data, randomly select w points--these are your Voronoi centers. A Voronoi cell encapsulates all neighboring points that are nearest to each center. Imagine if you assign a different color to each of Voronoi centers, so that each point assigned to a given center is painted that color. As long as you have a sufficient density, doing this will nicely show the boundaries of each Voronoi center (as the boundary that separates two colors.
How to select the Voronoi Centers? I use two orthogonal guidelines. After random selecting the w points, calculate the VT for your training data. Next check the number of data points assigned to each Voronoi center--these values should be about the same (given uniform point density across your data space). In two dimensions, this would cause a VT with tiles of the same size.That's the first rule, here's the second. Select w by iteration--run your kNN algorithm with w as a variable parameter, and measure performance (time required to return a prediction by querying the VT).
So imagine you have one million data points..... If the points were persisted in an ordinary 2D data structure, or in a kd-tree, you would perform on average a couple million distance calculations for each new data points whose response variable you wish to predict. Of course, those calculations are performed on a single data set. With a V/T, the nearest-neighbor search is performed in two steps one after the other, against two different populations of data--first against the Voronoi centers, then once the nearest center is found, the points inside the cell corresponding to that center are searched to find the actual nearest neighbor (by successive distance calculations) Combined, these two look-ups are much faster than a single brute-force look-up. That's easy to see: for 1M data points, suppose you select 250 Voronoi centers to tesselate your data space. On average, each Voronoi cell will have 4,000 data points. So instead of performing on average 500,000 distance calculations (brute force), you perform far lesss, on average just 125 + 2,000.
III. Calculating the Result (the predicted response variable)
There are two steps to calculating the predicted value from a set of kNN training data. The first is identifying n, or the number of nearest neighbors to use for this calculation. The second is how to weight their contribution to the predicted value.
W/r/t the first component, you can determine the best value of n by solving an optimization problem (very similar to least squares optimization). That's the theory; in practice, most people just use n=3. In any event, it's simple to run your kNN algorithm over a set of test instances (to calculate predicted values) for n=1, n=2, n=3, etc. and plot the error as a function of n. If you just want a plausible value for n to get started, again, just use n = 3.
The second component is how to weight the contribution of each of the neighbors (assuming n > 1).
The simplest weighting technique is just multiplying each neighbor by a weighting coefficient, which is just the 1/(dist * K), or the inverse of the distance from that neighbor to the test instance often multiplied by some empirically derived constant, K. I am not a fan of this technique because it often over-weights the closest neighbors (and concomitantly under-weights the more distant ones); the significance of this is that a given prediction can be almost entirely dependent on a single neighbor, which in turn increases the algorithm's sensitivity to noise.
A must better weighting function, which substantially avoids this limitation is the gaussian function, which in python, looks like this:
def weight_gauss(dist, sig=2.0) :
return math.e**(-dist**2/(2*sig**2))
To calculate a predicted value using your kNN code, you would identify the n nearest neighbors to the data point whose response variable you wish to predict ('test instance'), then call the weight_gauss function, once for each of the n neighbors, passing in the distance between each neighbor the the test point.This function will return the weight for each neighbor, which is then used as that neighbor's coefficient in the weighted average calculation.
What you are facing is known as the curse of dimensionality. It is sometimes useful to run an algorithm like PCA or ICA to make sure that you really need all 21 dimensions and possibly find a linear transformation which would allow you to use less than 21 with approximately the same result quality.
Update:
I encountered them in a book called Biomedical Signal Processing by Rangayyan (I hope I remember it correctly). ICA is not a trivial technique, but it was developed by researchers in Finland and I think Matlab code for it is publicly available for download. PCA is a more widely used technique and I believe you should be able to find its R or other software implementation. PCA is performed by solving linear equations iteratively. I've done it too long ago to remember how. = )
The idea is that you break up your signals into independent eigenvectors (discrete eigenfunctions, really) and their eigenvalues, 21 in your case. Each eigenvalue shows the amount of contribution each eigenfunction provides to each of your measurements. If an eigenvalue is tiny, you can very closely represent the signals without using its corresponding eigenfunction at all, and that's how you get rid of a dimension.
Top answers are good but old, so I'd like to add up a 2016 answer.
As said, in a high dimensional space, the curse of dimensionality lurks around the corner, making the traditional approaches, such as the popular k-d tree, to be as slow as a brute force approach. As a result, we turn our interest in Approximate Nearest Neighbor Search (ANNS), which in favor of some accuracy, speedups the process. You get a good approximation of the exact NN, with a good propability.
Hot topics that might be worthy:
Modern approaches of LSH, such as Razenshteyn's.
RKD forest: Forest(s) of Randomized k-d trees (RKD), as described in FLANN,
or in a more recent approach I was part of, kd-GeRaF.
LOPQ which stands for Locally Optimized Product Quantization, as described here. It is very similar to the new Babenko+Lemptitsky's approach.
You can also check my relevant answers:
Two sets of high dimensional points: Find the nearest neighbour in the other set
Comparison of the runtime of Nearest Neighbor queries on different data structures
PCL kd-tree implementation extremely slow
To answer your questions one by one:
No, euclidean distance is a bad metric in high dimensional space. Basically in high dimensions, data points have large differences between each other. That decreases the relative difference in the distance between a given data point and its nearest and farthest neighbour.
Lot of papers/research are there in high dimension data, but most of the stuff requires a lot of mathematical sophistication.
KD tree is bad for high dimensional data ... avoid it by all means
Here is a nice paper to get you started in the right direction. "When in Nearest Neighbour meaningful?" by Beyer et all.
I work with text data of dimensions 20K and above. If you want some text related advice, I might be able to help you out.
Cosine similarity is a common way to compare high-dimension vectors. Note that since it's a similarity not a distance, you'd want to maximize it not minimize it. You can also use a domain-specific way to compare the data, for example if your data was DNA sequences, you could use a sequence similarity that takes into account probabilities of mutations, etc.
The number of nearest neighbors to use varies depending on the type of data, how much noise there is, etc. There are no general rules, you just have to find what works best for your specific data and problem by trying all values within a range. People have an intuitive understanding that the more data there is, the fewer neighbors you need. In a hypothetical situation where you have all possible data, you only need to look for the single nearest neighbor to classify.
The k Nearest Neighbor method is known to be computationally expensive. It's one of the main reasons people turn to other algorithms like support vector machines.
kd-trees indeed won't work very well on high-dimensional data. Because the pruning step no longer helps a lot, as the closest edge - a 1 dimensional deviation - will almost always be smaller than the full-dimensional deviation to the known nearest neighbors.
But furthermore, kd-trees only work well with Lp norms for all I know, and there is the distance concentration effect that makes distance based algorithms degrade with increasing dimensionality.
For further information, you may want to read up on the curse of dimensionality, and the various variants of it (there is more than one side to it!)
I'm not convinced there is a lot use to just blindly approximating Euclidean nearest neighbors e.g. using LSH or random projections. It may be necessary to use a much more fine tuned distance function in the first place!
A lot depends on why you want to know the nearest neighbors. You might look into the mean shift algorithm http://en.wikipedia.org/wiki/Mean-shift if what you really want is to find the modes of your data set.
I think cosine on tf-idf of boolean features would work well for most problems. That's because its time-proven heuristic used in many search engines like Lucene. Euclidean distance in my experience shows bad results for any text-like data. Selecting different weights and k-examples can be done with training data and brute-force parameter selection.
iDistance is probably the best for exact knn retrieval in high-dimensional data. You can view it as an approximate Voronoi tessalation.
I've experienced the same problem and can say the following.
Euclidean distance is a good distance metric, however it's computationally more expensive than the Manhattan distance, and sometimes yields slightly poorer results, thus, I'd choose the later.
The value of k can be found empirically. You can try different values and check the resulting ROC curves or some other precision/recall measure in order to find an acceptable value.
Both Euclidean and Manhattan distances respect the Triangle inequality, thus you can use them in metric trees. Indeed, KD-trees have their performance severely degraded when the data have more than 10 dimensions (I've experienced that problem myself). I found VP-trees to be a better option.
KD Trees work fine for 21 dimensions, if you quit early,
after looking at say 5 % of all the points.
FLANN does this (and other speedups)
to match 128-dim SIFT vectors. (Unfortunately FLANN does only the Euclidean metric,
and the fast and solid
scipy.spatial.cKDTree
does only Lp metrics;
these may or may not be adequate for your data.)
There is of course a speed-accuracy tradeoff here.
(If you could describe your Ndata, Nquery, data distribution,
that might help people to try similar data.)
Added 26 April, run times for cKDTree with cutoff on my old mac ppc, to give a very rough idea of feasibility:
kdstats.py p=2 dim=21 N=1000000 nask=1000 nnear=2 cutoff=1000 eps=0 leafsize=10 clustype=uniformp
14 sec to build KDtree of 1000000 points
kdtree: 1000 queries looked at av 0.1 % of the 1000000 points, 0.31 % of 188315 boxes; better 0.0042 0.014 0.1 %
3.5 sec to query 1000 points
distances to 2 nearest: av 0.131 max 0.253
kdstats.py p=2 dim=21 N=1000000 nask=1000 nnear=2 cutoff=5000 eps=0 leafsize=10 clustype=uniformp
14 sec to build KDtree of 1000000 points
kdtree: 1000 queries looked at av 0.48 % of the 1000000 points, 1.1 % of 188315 boxes; better 0.0071 0.026 0.5 %
15 sec to query 1000 points
distances to 2 nearest: av 0.131 max 0.245
You could try a z order curve. It's easy for 3 dimension.
I had a similar question a while back. For fast Approximate Nearest Neighbor Search you can use the annoy library from spotify: https://github.com/spotify/annoy
This is some example code for the Python API, which is optimized in C++.
from annoy import AnnoyIndex
import random
f = 40
t = AnnoyIndex(f, 'angular') # Length of item vector that will be indexed
for i in range(1000):
v = [random.gauss(0, 1) for z in range(f)]
t.add_item(i, v)
t.build(10) # 10 trees
t.save('test.ann')
# ...
u = AnnoyIndex(f, 'angular')
u.load('test.ann') # super fast, will just mmap the file
print(u.get_nns_by_item(0, 1000)) # will find the 1000 nearest neighbors
They provide different distance measurements. Which distance measurement you want to apply depends highly on your individual problem. Also consider prescaling (meaning weighting) certain dimensions for importance first. Those dimension or feature importance weights might be calculated by something like entropy loss or if you have a supervised learning problem gini impurity gain or mean average loss, where you check how much worse your machine learning model performs, if you scramble this dimensions values.
Often the direction of the vector is more important than it's absolute value. For example in the semantic analysis of text documents, where we want document vectors to be close when their semantics are similar, not their lengths. Thus we can either normalize those vectors to unit length or use angular distance (i.e. cosine similarity) as a distance measurement.
Hope this is helpful.
Is Euclidean distance a good metric for finding the nearest neighbors in the first place? If not, what are my options?
I would suggest soft subspace clustering, a pretty common approach nowadays, where feature weights are calculated to find the most relevant dimensions. You can use these weights when using euclidean distance, for example. See curse of dimensionality for common problems and also this article can enlighten you somehow:
A k-means type clustering algorithm for subspace clustering of mixed numeric and
categorical datasets

Resources