prashant-jonny / accord Goto Github PK
View Code? Open in Web Editor NEWAutomatically exported from code.google.com/p/accord
Automatically exported from code.google.com/p/accord
What steps will reproduce the problem?
1. Decision tree with 6 continuous variables, 1 discrete
2. Learn the tree
3. Exception at C45Learning.cs:274
The reason for this is that the maxGainPartition array contains only one value,
and the algorithm expects two. Causes an IndexOutOfBounds exception.
It seems all partitions are assumed to be split in two but for some strange
reason my data produces one partition with only one subset.
What version of the product are you using? On what operating system?
2.8.1, Windows 7, 64-bit.
Attached is my training data in a flat format. First column is irrelevant, last
column is output (yes/no). All except the penultimate column are continuous
decision variables.
Original issue reported on code.google.com by [email protected]
on 11 Feb 2013 at 12:38
What steps will reproduce the problem?
1.
2.
3.
What is the expected output? What do you see instead?
What version of the product are you using? On what operating system?
v2.8.1
Please provide any additional information below.
i have downloaded the rar file. how to include these dlls in the project/how to
install them please help
Original issue reported on code.google.com by [email protected]
on 28 Feb 2013 at 4:25
User code contribution of Dynamic Time Warping by user david_r
Original issue reported on code.google.com by [email protected]
on 5 Apr 2012 at 8:44
Attachments:
Some classes from the Analysis namespace, such as LogisticRegressionAnalysis,
do not have source code examples.
Original issue reported on code.google.com by [email protected]
on 8 Nov 2012 at 12:19
It would be useful to have an implementation of the implicitly restarted
Lanczos method algorithm (http://en.wikipedia.org/wiki/Lanczos_algorithm). It
could improve the efficiency of methods which require only the largest or the
first largest Eigenvalues to be computed.
Original issue reported on code.google.com by [email protected]
on 14 Dec 2012 at 5:52
Here's code for Accord.MachineLearning.DecisionTrees.Learning.computeInfoGain:
{
threshold = 0;
if (tree.Attributes[attributeIndex].Nature == DecisionAttributeKind.Discrete)
return entropy - computeInfo(input, output, attributeIndex, out partitions);
return entropy - computeInfo(input, output, attributeIndex, out partitions, out threshold);
}
when the attribute is continous, split gain is SUBSTRACTED from entropy, but
the method that computes split gain for continous attributes (computeInfo)
already returns NEGATIVE value, so split gain should be ADDED to entropy.
Otherwise the c4.5 algorithm uses the worst of best possible splits.
so the last line should be as follows:
return entropy + computeInfo(input, output, attributeIndex, out partitions, out
threshold);
Original issue reported on code.google.com by [email protected]
on 9 Mar 2013 at 3:47
In order to use some models with categoric data, such as logistic regression,
it is first necessary to convert categories into factors. It would be nice to
have a more easier or standard way to create such variables.
Original issue reported on code.google.com by [email protected]
on 9 Dec 2012 at 7:20
What steps will reproduce the problem?
1. Try to traverse automatically a KDtree<double> using the foreach(...).
2. The output is empty.
What is the expected output? What do you see instead?
A list of KDTreeNode<double>.
What version of the product are you using? On what operating system?
2.8.1
Please provide any additional information below.
Looking at the source code, the condition stack.Count == 0 is wrong. it should
be stack.Count != 0.
Original issue reported on code.google.com by [email protected]
on 10 Jan 2013 at 11:56
What steps will reproduce the problem?
1. Use a SVM algorithm with SMO and Bootstrapping
2. At some point Bootstrapping can select all negative or all positive samples.
This breaks the code in
Accord.NET\Sources\Accord.MachineLearning\VectorMachines\Learning\SequentialMini
malOptimization.cs at line 479
3. It tries to find an upper and lower bound for the + en - samples but if
there are no + samples, an IndexOutOfRangeException is thrown here.
What is the expected output? What do you see instead?
I'm not too sure what the algorithm is supposed to do in this case. At least
catch the Exception?
What version of the product are you using? On what operating system?
Accord.NET v2.7.0 on Windows Vista
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 23 Jul 2012 at 3:13
Evaluate the feasibility of producing an Android compatible version using
Mono/Xamarin.
Original issue reported on code.google.com by [email protected]
on 14 Mar 2013 at 1:15
Hello,
I'm using Accord.NET framework to recognize some patterns using
MulticlassSupportVectorMachine class.
Here is my code:
Parallel.For(0, list.Length, i =>
{
// .. load some stuff ..
// here we get a NullReferenceException inside createCache() method.
int num = SupportVectorMachine.Compute(data, config.ComputeMethod, out responses);
if (!SupportVectorMachine.IsProbabilistic)
{
// Normalize responses
double max = responses.Max();
double min = responses.Min();
responses = Accord.Math.Tools.Scale(min, max, 0, 1, responses);
}
// .. do something ..
}
});
If I run this Parallel.For loop in a normal way (e.g using for or foreach), the
methods runs fine.
Can anyone help me, please?
Original issue reported on code.google.com by [email protected]
on 22 Oct 2012 at 7:49
Add support for probit regression using the already existent LinkFunctions.
Original issue reported on code.google.com by [email protected]
on 12 Feb 2013 at 1:25
What steps will reproduce the problem?
1. Variances and Degrees of Freedom a calculated from the attached data sets
Sample 1 (Random_50_0.5.txt) has a variance of 1.05766555271071 and DoF of 49.
and Sample 2 (Random_50_0.7.txt) has a variance of 1.16570834301777 and DoF of
49.
2. Enter data in Constructor FTest(Sample1.Variance, Sample2.Variance,
Sample1.DoF, Sample2.DoF, TwoSampleHypothesis.ValuesAreDifferent);
3. Compare these with One Tailed Tests using Constructors:
One Tailed - Greater:
FTest(Sample1.Variance, Sample2.Variance,
Sample1.DoF, Sample2.DoF, TwoSampleHypothesis.FirstValueIsGreaterThanSecond);
One Tailed - Less:
FTest(Sample1.Variance, Sample2.Variance,
Sample1.DoF, Sample2.DoF, TwoSampleHypothesis.FirstValueIsSmallerThanSecond));
What is the expected output? What do you see instead?
Results:
Two Tailed p-value: 1.26772493058357
One Tail Greater p-value: 0.633862465291784
One Tail Less p-value: 0.366137534708217
For the two tailed test I would have expected 0.732275069416434 (i.e. double
the One Tail Less, this value was confirmed by R and SciPy).
What version of the product are you using? On what operating system?
Windows8 (Visual Studio 2008), 64.bit, Accord.Net Version 2.8.0
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 20 Dec 2012 at 10:24
Attachments:
It would be interesting to have implementations of the AdaBoost learning
algorithm, specially if it could be used to learn Haar Cascade Classifiers
which could be used in the Vision namespace.
Original issue reported on code.google.com by [email protected]
on 19 Dec 2012 at 12:05
Please describe the new feature you are interested and how it would
contribute to the project.
Trimmed mean option parameter the Levene's Test. It is a common parameter of
the Levene's test in many statistical libraries.
Original issue reported on code.google.com by [email protected]
on 28 Feb 2013 at 3:16
What steps will reproduce the problem?
1.create an array of imputs for more than 10000 values in a row
2.more than 500 rows
3.SequentialMinimalOptimization
What is the expected output? What do you see instead?
out of memory exeption
Locating source for
'D:\Projects\Accord.NET\Sources\Accord.MachineLearning\VectorMachines\Learning\M
ulticlassSupportVectorLearning.cs'. Checksum: MD5 {56 e9 4d 58 59 a2 7e ab 57 5
af dc 91 88 72 a6}
The file
'D:\Projects\Accord.NET\Sources\Accord.MachineLearning\VectorMachines\Learning\M
ulticlassSupportVectorLearning.cs' does not exist.
Looking in script documents for
'D:\Projects\Accord.NET\Sources\Accord.MachineLearning\VectorMachines\Learning\M
ulticlassSupportVectorLearning.cs'...
Looking in the projects for
'D:\Projects\Accord.NET\Sources\Accord.MachineLearning\VectorMachines\Learning\M
ulticlassSupportVectorLearning.cs'.
The file was not found in a project.
Looking in directory 'C:\Program Files (x86)\Microsoft Visual Studio
11.0\Common7\IDE\vc7\atlmfc'...
Looking in directory 'C:\Program Files (x86)\Microsoft Visual Studio
11.0\Common7\IDE\vc7\crt'...
The debug source files settings for the active solution indicate that the
debugger will not ask the user to find the file:
D:\Projects\Accord.NET\Sources\Accord.MachineLearning\VectorMachines\Learning\Mu
lticlassSupportVectorLearning.cs.
The debugger could not locate the source file
'D:\Projects\Accord.NET\Sources\Accord.MachineLearning\VectorMachines\Learning\M
ulticlassSupportVectorLearning.cs'.
What version of the product are you using? On what operating system?
windows 7 .net 4.5 , accord version 2.7.1
Please provide any additional information below.
It only gives me an error when i add more than about 500 samples (inputs)Error
is from visual studio 12 express
Original issue reported on code.google.com by [email protected]
on 8 Nov 2012 at 7:00
need to change the indexing variable in the Translate method as follows:
public int[] Translate(params string[] data)
{
int[] result = new int[data.Length];
for (int i = 0; i < data.Length; i++)
{
for (int j = 0; j < Columns.Count; j++)
{
//Options options = this.Columns[i];
Options options = this.Columns[j]; // new index***
if (options.Mapping.TryGetValue(data[i], out result[i]))
break;
}
}
return result;
}
Thanks for sharing your project.... very helpful!
Scott
Original issue reported on code.google.com by [email protected]
on 21 Aug 2012 at 2:53
One-Way ANOVA doesn't works with different sample sizes. This is already fixed
on the development trunk and should be available on 2.7.1
Original issue reported on code.google.com by [email protected]
on 29 Aug 2012 at 3:53
What steps will reproduce the problem?
1.linear filtering
2.cross scale center-surround differences
3.winner-take-all networks
What is the expected output? What do you see instead?
saliency maps can be obtained.
What version of the product are you using? On what operating system?
I am using SaliencyToolbox2.2 which implemented by matlab code on widowns 7.
Please provide any additional information below.
[1] L.Itti, C.Koch, E. Niebur, A Model of Saliency-Based Visual Attention for
Rapid Scene Analysis, IEEE Transactions on Pattern Analysis and Machine
Intelligence 20(11) (1998)1254-1259.
[2] http://ilab.usc.edu/toolkit/
Original issue reported on code.google.com by [email protected]
on 26 Feb 2013 at 4:16
'''What steps will reproduce the problem?'''
* 1.Create two images with fully transparent areas that can be combined
* 2.the first image blends properly
* 3.the second images alpha channel is blended in. if its fully transparent
those bits should not copy
'''What is the expected output? What do you see instead?'''
100% transparent bits should not blend int
'''What version of the product are you using? On what operating system?'''
Accord 2.1.4
'''Please provide any additional information below:'''
Attached code fixes for 100% alpha, but doesn't attempt anything fancy for
partial transparency
{{{
// validate source pixel's coordinates
if ((ox >= 0) && (oy >= 0) && (ox < width) && (oy < height))
{
int c = oy * srcStride + ox * srcPixelSize;
// fill destination image with pixel from source image
if (srcPixelSize == 4 && src[c + 3] == 0)
{
// nothing to copy in
//NOP
}
else if (dst[3] > 0)
{
// there is a pixel from the other image here, blend
double d1 = distance(ox, oy, center1.X, center1.Y);
double d2 = distance(ox, oy, center2.X, center2.Y);
double f = Accord.Math.Tools.Scale(0, dmax, 0, 1, d1 - d2);
if (f < 0) f = 0;
if (f > 1) f = 1;
double f2 = (1.0 - f);
dst[0] = (byte)(src[c] * f2 + dst[0] * f);
}}}
Original issue reported on code.google.com by [email protected]
on 5 Apr 2012 at 8:46
Ensure BagOfWords and BagOfVisualWords are fully serializable.
Original issue reported on code.google.com by [email protected]
on 9 Dec 2012 at 7:17
These lines aren't needed (does nothing).
Original issue reported on code.google.com by [email protected]
on 18 Jul 2011 at 3:02
Attachments:
What steps will reproduce the problem?
1. Used the following data:
List<double[]> _samples = new List<double[]>();
double[] A = { 250, 260, 230, 270 };
double[] B = { 310, 330, 280, 360 };
double[] C = { 250, 230, 220, 260 };
double[] D = { 340, 270, 300, 320 };
double[] E = { 250, 240, 270, 290 };
_samples.Add(A);
_samples.Add(B);
_samples.Add(C);
_samples.Add(D);
_samples.Add(E);
LeveneTest result = new LeveneTest(_samples.ToArray(), true);
What is the expected output? What do you see instead?
Expect Output for result:
statistic= ~0.7247
pValue = ~0.5886
This was outputted by R (Library: lawstat, levene.test) and propriety
statistics software.
Actual Output for result:
statistic= ~0.7247
pValue = ~0.8228
What version of the product are you using? On what operating system?
2.8.1
Please provide any additional information below
Data Works fine for the newly added Bartlett's test
Original issue reported on code.google.com by [email protected]
on 3 Jan 2013 at 5:14
Add backward compatibility to the .NET Framework v3.5
Original issue reported on code.google.com by [email protected]
on 10 Sep 2012 at 11:55
What steps will reproduce the problem?
1.Start Visual C# 2010 and Open the Samples.sln installed in C:/Program
Files/Accord.Net/Framework/Samples and Launch the solver Project
2.The Form shows Objective Min 2x² + xy + y² - 5y.
Put the constraints as :
-x - 3y >= -2
-x - y >= 0
x >= 0
y >= 0
3.Click Compute
What is the expected output? What do you see instead?
the only feasible point is x=0,y=0 which should be the solution
the solution given by the Form is:
Solution:
2x² + xy + y² - 5y = 0.0947319461552687
x = -0.44079027173455
y = 0.44079027173455
which does not respect the constraint x>=0
What version of the product are you using? On what operating system?
I am using the version 2.6.1 of the library Accord.Net on windows XP and using
visual c# 2010
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 1 Jun 2012 at 3:45
ID3 (and possibly C4.5) generate a leaf node with Null output if a discrete
variable is said to have multiple possible values but some of the possible
values do not have any corresponding output.
It should be investigated if those algorithms handle this case and implement
the solution (use the mode?) appropriately.
Original issue reported on code.google.com by [email protected]
on 13 Feb 2013 at 2:22
Hi,
As far as I know (maybe it would need more research) the FREAK detector is more
powerfull, faster and consume less memory than SURF and SIFT.
Also, I don't think it is patented because in openCV it is not included in the
"nonfree" DLL.
It could be a very good alternative for users who want to use Accord.NET in
commercial application but can't because most of the very powerful algorithm
use SURF. All of this could be replaced by the FREAK detector.
Here is a comparison of 3 detector and FREAK is the winner:
http://computer-vision-talks.com/2012/08/a-battle-of-three-descriptors-surf-frea
k-and-brisk/
Original issue reported on code.google.com by [email protected]
on 1 Mar 2013 at 2:42
What steps will reproduce the problem?
Running a PLS-Regression with 3 dependent variables and 12 predictor variables,
and then computing the MultivariateLinearRegression obtained from the PLS.
What is the expected output? What do you see instead?
It should compute the regression coheficients correctly, instead it crashes
with a index out of range exception in the compute method.
What version of the product are you using? On what operating system?
v4.7 on Windows 7
Please provide any additional information below.
It seems you are taking the wrong dimensions in the compute function of the
MultivariateLinearRegression.
int M = coefficients.GetLength(1);
it should be:
int M = coefficients.GetLength(0);
Original issue reported on code.google.com by [email protected]
on 29 Jul 2012 at 6:01
It should be OK to add support for discrete-valued observations, but it will
also be necessary to make a notice that this may decrease performance.
Original issue reported on code.google.com by [email protected]
on 14 Jan 2013 at 10:42
What steps will reproduce the problem?
1. create the GoldfarbIdnaniQuadraticSolver with the following constraints
A B C D Op Val
0.0732 0.0799 0.1926 0.0047 = 0.098
1 1 1 1 = 1
1 0 0 0 >= 0
0 1 0 0 >= 0
0 0 1 0 >= 0
0 0 0 1 >= 0
1 0 0 0 >= 0.5
2. minimize with
Q:
0.12264004 0.011579293 0.103326825 0.064073439
0.011579293 0.033856 0.014311947 0.014732381
0.103326825 0.014311947 0.17715681 0.067615114
0.064073439 0.014732381 0.067615114 0.11539609
and
d:
0 0 0 0
What is the expected output? What do you see instead?
the expected results are
A B C D
0.5 0.336259542 0.163740458 0
The output are
A B C D
0.5 0.3996 0.1543 -0.0539
What version of the product are you using? On what operating system?
the version is 2.8.1.0 Last Published on 12/21/2012
OS is Window XP
Please provide any additional information below.
development Tools is Visual Studio 2010 Professional
more Properties from the solver:
Deletions: 2
Iterations: 7
Lagrangian[0]: 0.47610163141385442
Lagrangian[1]: 0.018357076997984511
Lagrangian[2]: 0.0
Lagrangian[3]: 0.0
Lagrangian[4]: 0.0
Lagrangian[5]: 0.057217556502738381
Lagrangian[6]: 0.057966723535120945
Solution[0]: 0.5
Solution[1]: 0.39959370168854663
Solution[2]: 0.15434219180400793
Solution[3]: -0.053935893492554586
Value: 0.03092710642619409
Original issue reported on code.google.com by [email protected]
on 24 Jan 2013 at 7:24
It would be interesting to add a feature extractor for computing histogram of
oriented gradients in the Accord.Imaging namespace.
Original issue reported on code.google.com by [email protected]
on 19 Dec 2012 at 12:02
KNearestNeighbor should not assume zero-based consecutive class labels. If the
labels (the outputs parameter of the constructor) happens to be something like
{4, 2, 5} (does not have zero and does not include all labels from 0 - 5), the
method Compute crashes.
Attached is a modified version of the class that works with this use case
without the problem.
Original issue reported on code.google.com by [email protected]
on 2 Nov 2012 at 12:14
Attachments:
What steps will reproduce the problem?
Issue linked from StackOverflow post:
http://stackoverflow.com/questions/13052167/using-accord-nets-codification-objec
t-to-codify-second-data-set/
What is the expected output? What do you see instead?
Expected Result is "lorem"
Please provide any additional information below.
Attached is a source code file which reproduces the issue.
It is entirely possible I am doing something wrong.
Original issue reported on code.google.com by [email protected]
on 25 Oct 2012 at 2:02
Attachments:
What steps will reproduce the problem?
1.
2.
3.
What is the expected output? What do you see instead?
What version of the product are you using? On what operating system?
Please provide any additional information below.
getting started gives the information about how to install executable installer
and not about how to install rar file. please help me. I am in urgent need of
it.
Original issue reported on code.google.com by [email protected]
on 2 Mar 2013 at 3:59
What steps will reproduce the problem?
1. I created an Id3Learning tree
2. After using id3Learning.Run() I attempted to serialize the tree.
What is the expected output? What do you see instead?
The program threw an exception stating that DecisionAttributeCollection is not
serializeable.
What version of the product are you using? On what operating system?
4.0.30319 on Windows 7 64 bit
Please provide any additional information below.
I was able to fix the error by adding the [Serializeable] flag to the
DecisionAttributeCollection class.
Original issue reported on code.google.com by [email protected]
on 5 Oct 2012 at 7:12
The C4.5 may end with a IndexOutOfRange after the 2.8.2 update.
Original issue reported on code.google.com by [email protected]
on 19 Mar 2013 at 4:32
The Binomial test yields an incorrect P-Value for the Binomial Two-Tailed test
but is correct for both One Tailed results compared to SciPy and R.
What steps will reproduce the problem?
1. use Constructor: BinomialTest(int successes, int trials, double
hypothesizedProbability, OneSampleHypothesis alternate);
2. No. of trials 235
3. No. of successes 51
4. HypothesizedProbability 1/6
What is the expected output? What do you see instead?
Two Tailed
Statistic = 51
Accord.Net = 1.96404518910672
R = 0.04375
SciPy= 0.043747970182413443
What version of the product are you using? On what operating system?
v2.8.2
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 28 Feb 2013 at 3:26
Add minimum neighbors threshold and rectangle averaging in the Haar object
detector.
Original issue reported on code.google.com by [email protected]
on 10 Sep 2012 at 11:54
I've already tried to submit issue but seems like it was timeout, so - once
again:
1. corner detector type - FastCornersDetector
2. feature detector type - FastRetinaKeypointDetector
3. point matcher type - KNearestNeighborMatching<byte[]>(2,
Distance.BitwiseHamming)
a bit of code:
cornerDetector.Threshold = 18;
List<FastRetinaKeypoint> descriptors = freakDetector.ProcessImage(newImage);
var union1 = descriptors.ToArray(); // length == 2, got it from newImage
(attached)
var union2 = lastFeaturePoints1; // length == 1, got it from oldImage (attached)
var knMatches = matcher.Match(union2, union1); // exception here
Exception details:
System.IndexOutOfRangeException was unhandled by user code
HResult=-2146233080
Message=Index was outside the bounds of the array.
Source=Accord.Math
StackTrace:
at Accord.Math.Matrix.Bottom[T](T[] values, Int32 count, Boolean inPlace) in C:\Work\Accord.NET\Sources\Accord.Math\Matrix\Matrix.Selection.cs:line 1657
at Accord.MachineLearning.KNearestNeighbors`1.Compute(T input, Double[]& scores) in C:\Work\Accord.NET\Sources\Accord.MachineLearning\KNearestNeighbors`1.cs:line 326
at Accord.MachineLearning.KNearestNeighbors`1.Compute(T input, Double& response) in C:\Work\Accord.NET\Sources\Accord.MachineLearning\KNearestNeighbors`1.cs:line 305
at Accord.Imaging.KNearestNeighborMatching`1.Match(IFeaturePoint`1[] points1, IFeaturePoint`1[] points2) in C:\Work\Accord.NET\Sources\Accord.Imaging\KNearestNeighborMatching.cs:line 128
at Exonya.EdgeDetector.Accord.MainWindow.ProcessImageFREAK(UnmanagedImage rgbImage) in c:\Work\Exonya.EdgeDetector1\Exonya.EdgeDetector.Accord\MainWindow.xaml.cs:line 274
at Exonya.EdgeDetector.Accord.MainWindow.video_NewFrame(Object sender, NewFrameEventArgs eventArgs) in c:\Work\Exonya.EdgeDetector1\Exonya.EdgeDetector.Accord\MainWindow.xaml.cs:line 130
at AForge.Video.DirectShow.VideoCaptureDevice.OnNewFrame(Bitmap image)
at AForge.Video.DirectShow.VideoCaptureDevice.Grabber.BufferCB(Double sampleTime, IntPtr buffer, Int32 bufferLen)
InnerException:
Windows 7, Accord.NET 2.9.0.0
thanks
Original issue reported on code.google.com by [email protected]
on 17 May 2013 at 10:12
Attachments:
The learning of Logistic Regression models is done in the same fashion as
AForge.NET's Neural Networks: a Run method has to be called to run iterations
(or batch of iterations) manually, until the user decides convergence has been
reached.
The Cox's models, the Hidden Markov Models and Hidden Conditional Random
Fields, on the other hand, control the iteration themselves. You specify the
number of iterations and tolerance threshold and call them once.
It is needed to decide which is most appropriated and enforce this as a common
standard for all learning methods.
Original issue reported on code.google.com by [email protected]
on 7 Nov 2012 at 11:07
What steps will reproduce the problem?
1. Use HaarObjectDetector class to detect objects in an image
2. Set the MinSize property to something larger than the default
3. Observe that rectangles returned by the ProcessFrame method can be smaller
than the value of MinSize
What is the expected output? What do you see instead?
I would expect that only rectangles (indicating objects discovered) that are
larger than MinSize are returned by ProcessFrame.
What version of the product are you using? On what operating system?
2.8.1 on both Windows 7 & Windows 8
Please provide any additional information below.
Changing line 448 of HaarObjectDetector.cs from:
"if (window.Width < minSize.Width && window.Height < minSize.Height &&
window.Width > maxSize.Width && window.Height > maxSize.Height)"
to:
"if ((window.Width < minSize.Width || window.Height < minSize.Height))"
Solves the issue for me, although I'm not sure if this is in line with the
original intent
Original issue reported on code.google.com by [email protected]
on 25 Jan 2013 at 1:56
//Some description please
Original issue reported on code.google.com by [email protected]
on 11 Apr 2013 at 12:25
Attachments:
What steps will reproduce the problem?
1. Instead of example.xls file i want to attach my excel file to the SVM
Classifier, When i try to attach my file (attached), the problem i had faced is
the dataTable object is unable to typecast or read it properly to store in type
double Matrix. Please suggest me is there any specific file format for
example.xls as both sheet include 3 column X, Y and G.
What is the expected output? What do you see instead?
I need that SVM will read a generic xls file, with a specified file format of
column.
What version of the product are you using? On what operating system?
Window 7 , Latest Version.
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 12 May 2013 at 12:28
Attachments:
I think Transform method should be like this.
Original issue reported on code.google.com by [email protected]
on 17 Jul 2011 at 11:06
Attachments:
What steps will reproduce the problem?
1.Install Accord & AForge
2.Create project and add reference to the Accord.dll
3. The interfaces aren't recognized the the version information is 0.0.0
What is the expected output? What do you see instead?
It should show the versioning information in the reference. Instead I get a
message the the file cannot be referenced
What version of the product are you using? On what operating system?
Win 7
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 21 Nov 2012 at 10:01
Add examples for the PCA/LDA analysis using images.
Original issue reported on code.google.com by [email protected]
on 10 Jan 2013 at 11:50
What steps will reproduce the problem?
1. Create a new GridSearch object with 2*2 parameters
2. Specify a Fitting function throwing a ConvergenceException once. Other calls
return a normal double value e.g [0.2, 0.5, 0.1]
3. Call Compute() on the GridSearch object
What is the expected output? What do you see instead?
As the best error rate in the received GridSearchResult 0.1 is expected.
Instead, the Error property of the result is set to double.NaN
What version of the product are you using? On what operating system?
Accord 2.2.8
Please provide any additional information below.
double.NaN is set as error rate if a ConvergenceException is thrown while
computing the Fitting function. This error rate is added to the errors array.
At the end, the best error rate is estimated by the Min function of the System
Library. If a value in the error array is set to double.NaN, this rate is
selected as the Min value
var doubles = new[] {0.5, double.NaN, 1.0, 3.0};
Console.WriteLine(doubles.Min());
//writes "n. Def"
Original issue reported on code.google.com by [email protected]
on 23 May 2013 at 11:16
Two small fixes: removed dependence on Accord.Audio.Formats (as it is part of
Accord.Audio) and fixed the package name to Accord.DirectAudio. Without these
fixes, package cannot be installed via NuGet.
Original issue reported on code.google.com by [email protected]
on 1 Dec 2012 at 11:50
Attachments:
[http://accord-net.origo.ethz.ch/node/184 Dan originally wrote]:
In the correlationMatching.cs, there is a formula for Normalized Cross
Correlation matrix like:
for (int i = 0; i < windowSize; i++)
{
for (int j = 0; j < windowSize; j++)
{
sum1 += w1[i, j] * w2[i, j];
sum2 += w2[i, j] * w2[i, j];
}
}
matrix[n1, n2] = sum1 / System.Math.Sqrt(sum2);
But in reference pages, I found that sum2 should be like : sum2 += w1[i, j] *
w1[i, j] * w2[i, j] * w2[i, j];
Original issue reported on code.google.com by [email protected]
on 5 Apr 2012 at 8:50
What steps will reproduce the problem?
1.
2.
3.
What is the expected output? What do you see instead?
The SVM problem when used for Regression set, it has some problem
What version of the product are you using? On what operating system?
window 7
Please provide any additional information below.
Original issue reported on code.google.com by [email protected]
on 6 Jul 2012 at 2:41
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.