这里是libsvm的FQA :

官网请访问：http://www.csie.ntu.edu.tw/~cjlin/libsvm/faq.html#_TOP

LIBSVM FAQ

last modified : Tue, 20 Oct 2015 13:43:40 GMT

All Questions(84)

Q01:_Some_sample_uses_of_libsvm(2)
Q02:_Installation_and_running_the_program(13)
Q03:_Data_preparation(7)
Q04:_Training_and_prediction(29)
Q05:_Cross_validation_and_parameter_selection(9)
Q06:_Probability_outputs(3)
Q07:_Graphic_interface(3)
Q08:_Java_version_of_libsvm(4)
Q09:_Python_interface(1)
Q10:_MATLAB_OCTAVE_interface(13)

Q: Some courses which have used libsvm as a tool

[Go Top]

Q: Some applications/tools which have used libsvm
(and maybe liblinear).

LIBPMK: A Pyramid Match Toolkit
Maltparser: a system for data-driven dependency parsing
PyMVPA: python tool for classifying neuroimages
SOLpro: protein solubility predictor
BDVal: biomarker discovery in high-throughput datasets.
Realtime object recognition
scikits.learn: machine learning in Python

[Go Top]

Q: Where can I find documents/videos of libsvm ?

Official implementation document:
C.-C. Chang and C.-J. Lin. LIBSVM : a library for support vector machines. ACM Transactions on Intelligent Systems and Technology, 2:27:1--27:27, 2011. pdf, ps.gz, ACM digital lib.
Instructions for using LIBSVM are in the README files in the main directory and some sub-directories.
README in the main directory: details all options, data format, and library calls.
tools/README: parameter selection and other tools
A guide for beginners:
C.-W. Hsu, C.-C. Chang, and C.-J. Lin. A practical guide to support vector classification
An introductory video for windows users.

[Go Top]

Q: Where are change log and earlier versions?

See the change log.

You can download earlier versions here.

[Go Top]

Q: How to cite LIBSVM?

Please cite the following paper:

Chih-Chung Chang and Chih-Jen Lin, LIBSVM : a library for support vector machines. ACM Transactions on Intelligent Systems and Technology, 2:27:1--27:27, 2011. Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm

The bibtex format is

@article{CC01a,
 author = {Chang, Chih-Chung and Lin, Chih-Jen},
 title = {{LIBSVM}: A library for support vector machines},
 journal = {ACM Transactions on Intelligent Systems and Technology},
 volume = {2},
 issue = {3},
 year = {2011},
 pages = {27:1--27:27},
 note =	 {Software available at \url{http://www.csie.ntu.edu.tw/~cjlin/libsvm}}
}

[Go Top]

Q: I would like to use libsvm in my software. Is there any license problem?

We have "the modified BSD license," so it is very easy to use libsvm in your software. Please check the COPYRIGHT file in detail. Basically you need to

Clearly indicate that LIBSVM is used.
Retain the LIBSVM COPYRIGHT file in your software.

It can also be used in commercial products.

[Go Top]

Q: Is there a repository of additional tools based on libsvm?

Yes, see libsvm tools

[Go Top]

Q: On unix machines, I got "error in loading shared libraries" or "cannot open shared object file." What happened ?

This usually happens if you compile the code on one machine and run it on another which has incompatible libraries. Try to recompile the program on that machine or use static linking.

[Go Top]

Q: I have modified the source and would like to build the graphic interface "svm-toy" on MS windows. How should I do it ?

Build it as a project by choosing "Win32 Project." On the other hand, for "svm-train" and "svm-predict" you want to choose "Win32 Console Project." After libsvm 2.5, you can also use the file Makefile.win. See details in README.

If you are not using Makefile.win and see the following link error

LIBCMTD.lib(wwincrt0.obj) : error LNK2001: unresolved external symbol
_wWinMain@16

you may have selected a wrong project type.

[Go Top]

Q: I am an MS windows user but why only one (svm-toy) of those precompiled .exe actually runs ?

You need to open a command window and type svmtrain.exe to see all options. Some examples are in README file.

[Go Top]

Q: What is the difference between "." and "*" outputed during training?

"." means every 1,000 iterations (or every #data iterations is your #data is less than 1,000). "*" means that after iterations of using a smaller shrunk problem, we reset to use the whole set. See the implementation document for details.

[Go Top]

Q: Why occasionally the program (including MATLAB or other interfaces) crashes and gives a segmentation fault?

Very likely the program consumes too much memory than what the operating system can provide. Try a smaller data and see if the program still crashes.

[Go Top]

Q: How to build a dynamic library (.dll file) on MS windows?

The easiest way is to use Makefile.win. See details in README. Alternatively, you can use Visual C++. Here is the example using Visual Studio 2013:

Create a Win32 empty DLL project and set (in Project->$Project_Name Properties...->Configuration) to "Release." About how to create a new dynamic link library, please refer to http://msdn2.microsoft.com/en-us/library/ms235636(VS.80).aspx
Add svm.cpp, svm.h to your project.
Add __WIN32__ and _CRT_SECURE_NO_DEPRECATE to Preprocessor definitions (in Project->$Project_Name Properties...->C/C++->Preprocessor)
Set Create/Use Precompiled Header to Not Using Precompiled Headers (in Project->$Project_Name Properties...->C/C++->Precompiled Headers)
Set the path for the Modulation Definition File svm.def (in Project->$Project_Name Properties...->Linker->input
Build the DLL.
Rename the dll file to libsvm.dll and move it to the correct path.

[Go Top]

Q: On some systems (e.g., Ubuntu), compiling LIBSVM gives many warning messages. Is this a problem and how to disable the warning message?

If you are using a version before 3.18, probably you see a warning message like

svm.cpp:2730: warning: ignoring return value of int fscanf(FILE*, const char*, ...), declared with attribute warn_unused_result

This is not a problem; see this page for more details of ubuntu systems. To disable the warning message you can replace

CFLAGS = -Wall -Wconversion -O3 -fPIC

with

CFLAGS = -Wall -Wconversion -O3 -fPIC -U_FORTIFY_SOURCE

in Makefile.

After version 3.18, we have a better setting so that such warning messages do not appear.

[Go Top]

Q: In LIBSVM, why you don't use certain C/C++ library functions to make the code shorter?

For portability, we use only features defined in ISO C89. Note that features in ISO C99 may not be available everywhere. Even the newest gcc lacks some features in C99 (see http://gcc.gnu.org/c99status.html for details). If the situation changes in the future, we might consider using these newer features.

[Go Top]

Q: Why sometimes not all attributes of a data appear in the training/model files ?

libsvm uses the so called "sparse" format where zero values do not need to be stored. Hence a data with attributes

1 0 2 0

is represented as

1:1 3:2

[Go Top]

Q: What if my data are non-numerical ?

Currently libsvm supports only numerical data. You may have to change non-numerical data to numerical. For example, you can use several binary attributes to represent a categorical attribute.

[Go Top]

Q: Why do you consider sparse format ? Will the training of dense data be much slower ?

This is a controversial issue. The kernel evaluation (i.e. inner product) of sparse vectors is slower so the total training time can be at least twice or three times of that using the dense format. However, we cannot support only dense format as then we CANNOT handle extremely sparse cases. Simplicity of the code is another concern. Right now we decide to support the sparse format only.

[Go Top]

Q: Why sometimes the last line of my data is not read by svm-train?

We assume that you have '\n' in the end of each line. So please press enter in the end of your last line.

[Go Top]

Q: Is there a program to check if my data are in the correct format?

The svm-train program in libsvm conducts only a simple check of the input data. To do a detailed check, after libsvm 2.85, you can use the python script tools/checkdata.py. See tools/README for details.

[Go Top]

Q: May I put comments in data files?

We don't officially support this. But, currently LIBSVM is able to process data in the following format:

1 1:2 2:1 # your comments

Note that the character ":" should not appear in your comments.

[Go Top]

Q: How to convert other data formats to LIBSVM format?

It depends on your data format. A simple way is to use libsvmwrite in the libsvm matlab/octave interface. Take a CSV (comma-separated values) file in UCI machine learning repository as an example. We download SPECTF.train. Labels are in the first column. The following steps produce a file in the libsvm format.

matlab> SPECTF = csvread('SPECTF.train'); % read a csv file
matlab> labels = SPECTF(:, 1); % labels from the 1st column
matlab> features = SPECTF(:, 2:end); 
matlab> features_sparse = sparse(features); % features must be in a sparse matrix
matlab> libsvmwrite('SPECTFlibsvm.train', labels, features_sparse);

The tranformed data are stored in SPECTFlibsvm.train.

Alternatively, you can use convert.c to convert CSV format to libsvm format.

[Go Top]

Q: The output of training C-SVM is like the following. What do they mean?

optimization finished, #iter = 219
nu = 0.431030
obj = -100.877286, rho = 0.424632
nSV = 132, nBSV = 107
Total nSV = 132

obj is the optimal objective value of the dual SVM problem. rho is the bias term in the decision function sgn(w^Tx - rho). nSV and nBSV are number of support vectors and bounded support vectors (i.e., alpha_i = C). nu-svm is a somewhat equivalent form of C-SVM where C is replaced by nu. nu simply shows the corresponding parameter. More details are in libsvm document.

[Go Top]

Q: Can you explain more about the model file?

In the model file, after parameters and other informations such as labels , each line represents a support vector. Support vectors are listed in the order of "labels" shown earlier. (i.e., those from the first class in the "labels" list are grouped first, and so on.) If k is the total number of classes, in front of a support vector in class j, there are k-1 coefficients y*alpha where alpha are dual solution of the following two class problems:
1 vs j, 2 vs j, ..., j-1 vs j, j vs j+1, j vs j+2, ..., j vs k
and y=1 in first j-1 coefficients, y=-1 in the remaining k-j coefficients. For example, if there are 4 classes, the file looks like:

+-+-+-+--------------------+
|1|1|1|                    |
|v|v|v|  SVs from class 1  |
|2|3|4|                    |
+-+-+-+--------------------+
|1|2|2|                    |
|v|v|v|  SVs from class 2  |
|2|3|4|                    |
+-+-+-+--------------------+
|1|2|3|                    |
|v|v|v|  SVs from class 3  |
|3|3|4|                    |
+-+-+-+--------------------+
|1|2|3|                    |
|v|v|v|  SVs from class 4  |
|4|4|4|                    |
+-+-+-+--------------------+

[Go Top]

Q: Should I use float or double to store numbers in the cache ?

We have float as the default as you can store more numbers in the cache. In general this is good enough but for few difficult cases (e.g. C very very large) where solutions are huge numbers, it might be possible that the numerical precision is not enough using only float.

[Go Top]

Q: Does libsvm have special treatments for linear SVM?

No, libsvm solves linear/nonlinear SVMs by the same way. Some tricks may save training/testing time if the linear kernel is used, so libsvm is NOT particularly efficient for linear SVM, especially when C is large and the number of data is much larger than the number of attributes. You can either

Use small C only. We have shown in the following paper that after C is larger than a certain threshold, the decision function is the same.
S. S. Keerthi and C.-J. Lin. Asymptotic behaviors of support vector machines with Gaussian kernel . Neural Computation, 15(2003), 1667-1689.
Check liblinear, which is designed for large-scale linear classification.

Please also see our SVM guide on the discussion of using RBF and linear kernels.

[Go Top]

Q: The number of free support vectors is large. What should I do?

This usually happens when the data are overfitted. If attributes of your data are in large ranges, try to scale them. Then the region of appropriate parameters may be larger. Note that there is a scale program in libsvm.

[Go Top]

Q: Should I scale training and testing data in a similar way?

Yes, you can do the following:

> svm-scale -s scaling_parameters train_data > scaled_train_data
> svm-scale -r scaling_parameters test_data > scaled_test_data

[Go Top]

Q: On windows sometimes svm-scale.exe generates some non-ASCII data not good for training/prediction?

In general this does not happen, but we have observed in some rare situations, the output of svm-scale.exe directed to a file (by ">") has wrong encoding. That is, the file is not an ASCII file, so cannot be used for training/prediction. Please let us know if this happens as at this moment we don't clearly see how to fix the problem.

[Go Top]

Q: Does it make a big difference if I scale each attribute to [0,1] instead of [-1,1]?

For the linear scaling method, if the RBF kernel is used and parameter selection is conducted, there is no difference. Assume Mi and mi are respectively the maximal and minimal values of the ith attribute. Scaling to [0,1] means

                x'=(x-mi)/(Mi-mi)

For [-1,1],

                x''=2(x-mi)/(Mi-mi)-1.

In the RBF kernel,

                x'-y'=(x-y)/(Mi-mi), x''-y''=2(x-y)/(Mi-mi).

Hence, using (C,g) on the [0,1]-scaled data is the same as (C,g/2) on the [-1,1]-scaled data.

Though the performance is the same, the computational time may be different. For data with many zero entries, [0,1]-scaling keeps the sparsity of input data and hence may save the time.

[Go Top]

Q: The prediction rate is low. How could I improve it?

Try to use the model selection tool grid.py in the tools directory find out good parameters. To see the importance of model selection, please see our guide for beginners: A practical guide to support vector classification

[Go Top]

Q: My data are unbalanced. Could libsvm handle such problems?

Yes, there is a -wi options. For example, if you use

> svm-train -s 0 -c 10 -w1 1 -w-1 5 data_file

the penalty for class "-1" is larger. Note that this -w option is for C-SVC only.

[Go Top]

Q: What is the difference between nu-SVC and C-SVC?

Basically they are the same thing but with different parameters. The range of C is from zero to infinity but nu is always between [0,1]. A nice property of nu is that it is related to the ratio of support vectors and the ratio of the training error.

[Go Top]

Q: The program keeps running (without showing any output). What should I do?

You may want to check your data. Each training/testing data must be in one line. It cannot be separated. In addition, you have to remove empty lines.

[Go Top]

Q: The program keeps running (with output, i.e. many dots). What should I do?

In theory libsvm guarantees to converge. Therefore, this means you are handling ill-conditioned situations (e.g. too large/small parameters) so numerical difficulties occur.

You may get better numerical stability by replacing

typedef float Qfloat;

in svm.cpp with

typedef double Qfloat;

That is, elements in the kernel cache are stored in double instead of single. However, this means fewer elements can be put in the kernel cache.

[Go Top]

Q: The training time is too long. What should I do?

For large problems, please specify enough cache size (i.e., -m). Slow convergence may happen for some difficult cases (e.g. -c is large). You can try to use a looser stopping tolerance with -e. If that still doesn't work, you may train only a subset of the data. You can use the program subset.py in the directory "tools" to obtain a random subset.

If you have extremely large data and face this difficulty, please contact us. We will be happy to discuss possible solutions.

When using large -e, you may want to check if -h 0 (no shrinking) or -h 1 (shrinking) is faster. See a related question below.

[Go Top]

Q: Does shrinking always help?

If the number of iterations is high, then shrinking often helps. However, if the number of iterations is small (e.g., you specify a large -e), then probably using -h 0 (no shrinking) is better. See the implementation document for details.

[Go Top]

Q: How do I get the decision value(s)?

We print out decision values for regression. For classification, we solve several binary SVMs for multi-class cases. You can obtain values by easily calling the subroutine svm_predict_values. Their corresponding labels can be obtained from svm_get_labels. Details are in README of libsvm package.

If you are using MATLAB/OCTAVE interface, svmpredict can directly give you decision values. Please see matlab/README for details.

We do not recommend the following. But if you would like to get values for TWO-class classification with labels +1 and -1 (note: +1 and -1 but not things like 5 and 10) in the easiest way, simply add

		printf("%f\n", dec_values[0]*model->label[0]);

after the line

		svm_predict_values(model, x, dec_values);

of the file svm.cpp. Positive (negative) decision values correspond to data predicted as +1 (-1).

[Go Top]

Q: How do I get the distance between a point and the hyperplane?

The distance is |decision_value| / |w|. We have |w|^2 = w^Tw = alpha^T Q alpha = 2*(dual_obj + sum alpha_i). Thus in svm.cpp please find the place where we calculate the dual objective value (i.e., the subroutine Solve()) and add a statement to print w^Tw. More precisely, here is what you need to do

Search for "calculate objective value" in svm.cpp
In that place, si->obj is the variable for the objective value
Add a for loop to calculate the sum of alpha
Calculate 2*(si->obj + sum of alpha) and print the square root of it. You now get |w|. You need to recompile the code
Check an earlier FAQ on printing decision values. You need to recompile the code
Then print decision value divided by the |w| value obtained earlier.

[Go Top]

Q: On 32-bit machines, if I use a large cache (i.e. large -m) on a linux machine, why sometimes I get "segmentation fault ?"

On 32-bit machines, the maximum addressable memory is 4GB. The Linux kernel uses 3:1 split which means user space is 3G and kernel space is 1G. Although there are 3G user space, the maximum dynamic allocation memory is 2G. So, if you specify -m near 2G, the memory will be exhausted. And svm-train will fail when it asks more memory. For more details, please read this article.

The easiest solution is to switch to a 64-bit machine. Otherwise, there are two ways to solve this. If your machine supports Intel's PAE (Physical Address Extension), you can turn on the option HIGHMEM64G in Linux kernel which uses 4G:4G split for kernel and user space. If you don't, you can try a software `tub' which can eliminate the 2G boundary for dynamic allocated memory. The `tub' is available at http://www.bitwagon.com/tub.html.

[Go Top]

Q: How do I disable screen output of svm-train?

For commend-line users, use the -q option:

> ./svm-train -q heart_scale

For library users, set the global variable

extern void (*svm_print_string) (const char *);

to specify the output format. You can disable the output by the following steps:

Declare a function to output nothing:
```
void print_null(const char *s) {}
```
Assign the output function of libsvm by
```
svm_print_string = &print_null;
```

Finally, a way used in earlier libsvm is by updating svm.cpp from

#if 1
void info(const char *fmt,...)

#if 0
void info(const char *fmt,...)

[Go Top]

Q: I would like to use my own kernel. Any example? In svm.cpp, there are two subroutines for kernel evaluations: k_function() and kernel_function(). Which one should I modify ?

An example is "LIBSVM for string data" in LIBSVM Tools.

The reason why we have two functions is as follows. For the RBF kernel exp(-g |xi - xj|^2), if we calculate xi - xj first and then the norm square, there are 3n operations. Thus we consider exp(-g (|xi|^2 - 2dot(xi,xj) +|xj|^2)) and by calculating all |xi|^2 in the beginning, the number of operations is reduced to 2n. This is for the training. For prediction we cannot do this so a regular subroutine using that 3n operations is needed. The easiest way to have your own kernel is to put the same code in these two subroutines by replacing any kernel.

[Go Top]

Q: What method does libsvm use for multi-class SVM ? Why don't you use the "1-against-the rest" method?

It is one-against-one. We chose it after doing the following comparison: C.-W. Hsu and C.-J. Lin. A comparison of methods for multi-class support vector machines , IEEE Transactions on Neural Networks, 13(2002), 415-425.

"1-against-the rest" is a good method whose performance is comparable to "1-against-1." We do the latter simply because its training time is shorter.

[Go Top]

Q: I would like to solve L2-loss SVM (i.e., error term is quadratic). How should I modify the code ?

It is extremely easy. Taking c-svc for example, to solve

min_w w^Tw/2 + C \sum max(0, 1- (y_i w^Tx_i+b))^2,

only two places of svm.cpp have to be changed. First, modify the following line of solve_c_svc from

	s.Solve(l, SVC_Q(*prob,*param,y), minus_ones, y,
		alpha, Cp, Cn, param->eps, si, param->shrinking);

	s.Solve(l, SVC_Q(*prob,*param,y), minus_ones, y,
		alpha, INF, INF, param->eps, si, param->shrinking);

Second, in the class of SVC_Q, declare C as a private variable:

	double C;

In the constructor replace

	for(int i=0;i<prob.l;i++)
		QD[i]= (Qfloat)(this->*kernel_function)(i,i);

with

        this->C = param.C;
	for(int i=0;i<prob.l;i++)
		QD[i]= (Qfloat)(this->*kernel_function)(i,i)+0.5/C;

Then in the subroutine get_Q, after the for loop, add

        if(i >= start && i < len) 
		data[i] += 0.5/C;

For one-class svm, the modification is exactly the same. For SVR, you don't need an if statement like the above. Instead, you only need a simple assignment:

	data[real_i] += 0.5/C;

For large linear L2-loss SVM, please use LIBLINEAR.

[Go Top]

Q: In one-class SVM, parameter nu should be an upper bound of the training error rate. Why sometimes I get a training error rate bigger than nu?

At optimum, some training instances should satisfy w^Tx - rho = 0. However, numerically they may be slightly smaller than zero Then they are wrongly counted as training errors. You can use a smaller stopping tolerance (by the -e option) to make this problem less serious.

This issue does not occur for nu-SVC for two-class classification. We have that

nu is an upper bound on the ratio of training points on the wrong side of the hyperplane, and
therefore, nu is also an upper bound on the training error rate.

Numerical issues occur in calculating the first case because some training points satisfying y(w^Tx + b) - rho = 0 become negative. However, we have no numerical problems for the second case because we compare y(w^Tx + b) and 0 for counting training errors.

[Go Top]

Q: Why the code gives NaN (not a number) results?

This rarely happens, but few users reported the problem. It seems that their computers for training libsvm have the VPN client running. The VPN software has some bugs and causes this problem. Please try to close or disconnect the VPN client.

[Go Top]

Q: Why the sign of predicted labels and decision values are sometimes reversed?

This situation may occur before version 3.17. Nothing is wrong. Very likely you have two labels +1/-1 and the first instance in your data has -1. We give the following explanation.

Internally class labels are ordered by their first occurrence in the training set. For a k-class data, internally labels are 0, ..., k-1, and each two-class SVM considers pair (i, j) with i < j. Then class i is treated as positive (+1) and j as negative (-1). For example, if the data set has labels +5/+10 and +10 appears first, then internally the +5 versus +10 SVM problem has +10 as positive (+1) and +5 as negative (-1).

By this setting, if you have labels +1 and -1, it's possible that internally they correspond to -1 and +1, respectively. Some new users have been confused about this, so after version 3.17, if the data set has only two labels +1 and -1, internally we ensure +1 to be before -1. Then class +1 is always treated as positive in the SVM problem. Note that this is for two-class data only.

[Go Top]

Q: I don't know class labels of test data. What should I put in the first column of the test file?

Any value is ok. In this situation, what you will use is the output file of svm-predict, which gives predicted class labels.

[Go Top]

Q: How can I use OpenMP to parallelize LIBSVM on a multicore/shared-memory computer?

It is very easy if you are using GCC 4.2 or after.

In Makefile, add -fopenmp to CFLAGS.

In class SVC_Q of svm.cpp, modify the for loop of get_Q to:

#pragma omp parallel for private(j) schedule(guided)
			for(j=start;j<len;j++)

In the subroutine svm_predict_values of svm.cpp, add one line to the for loop:

#pragma omp parallel for private(i) schedule(guided) 
		for(i=0;i<l;i++)
			kvalue[i] = Kernel::k_function(x,model->SV[i],model->param);

For regression, you need to modify class SVR_Q instead. The loop in svm_predict_values is also different because you need a reduction clause for the variable sum:

#pragma omp parallel for private(i) reduction(+:sum) schedule(guided)
		for(i=0;i<model->l;i++)
			sum += sv_coef[i] * Kernel::k_function(x,model->SV[i],model->param);

Then rebuild the package. Kernel evaluations in training/testing will be parallelized. An example of running this modification on an 8-core machine using the data set real-sim:

8 cores:

%setenv OMP_NUM_THREADS 8
%time svm-train -c 8 -g 0.5 -m 1000 real-sim
175.90sec

1 core:

%setenv OMP_NUM_THREADS 1
%time svm-train -c 8 -g 0.5 -m 1000 real-sim
588.89sec

For this data, kernel evaluations take 91% of training time. In the above example, we assume you use csh. For bash, use

export OMP_NUM_THREADS=8

instead.

For Python interface, you need to add the -lgomp link option:

$(CXX) -lgomp -shared -dynamiclib svm.o -o libsvm.so.$(SHVER)

For MS Windows, you need to add /openmp in CFLAGS of Makefile.win

[Go Top]

Q: How could I know which training instances are support vectors?

It's very simple. Since version 3.13, you can use the function

void svm_get_sv_indices(const struct svm_model *model, int *sv_indices)

to get indices of support vectors. For example, in svm-train.c, after

		model = svm_train(&prob, &param);

you can add

		int nr_sv = svm_get_nr_sv(model);
		int *sv_indices = Malloc(int, nr_sv);
		svm_get_sv_indices(model, sv_indices);
		for (int i=0; i<nr_sv; i++)
			printf("instance %d is a support vector\n", sv_indices[i]);

If you use matlab interface, you can directly check

model.sv_indices

[Go Top]

Q: Why sv_indices (indices of support vectors) are not stored in the saved model file?

Although sv_indices is a member of the model structure to indicate support vectors in the training set, we do not store its contents in the model file. The model file is mainly used in the future for prediction, so it is basically independent from training data. Thus storing sv_indices is not necessary. Users should find support vectors right after the training process. See the previous FAQ.

[Go Top]

Q: After doing cross validation, why there is no model file outputted ?

Cross validation is used for selecting good parameters. After finding them, you want to re-train the whole data without the -v option.

[Go Top]

Q: Why my cross-validation results are different from those in the Practical Guide?

Due to random partitions of the data, on different systems CV accuracy values may be different.

[Go Top]

Q: On some systems CV accuracy is the same in several runs. How could I use different data partitions? In other words, how do I set random seed in LIBSVM?

If you use GNU C library, the default seed 1 is considered. Thus you always get the same result of running svm-train -v. To have different seeds, you can add the following code in svm-train.c:

#include <time.h>

and in the beginning of main(),

srand(time(0));

Alternatively, if you are not using GNU C library and would like to use a fixed seed, you can have

srand(1);

For Java, the random number generator is initialized using the time information. So results of two CV runs are different. To fix the seed, after version 3.1 (released in mid 2011), you can add

svm.rand.setSeed(0);

in the main() function of svm_train.java.

If you use CV to select parameters, it is recommended to use identical folds under different parameters. In this case, you can consider fixing the seed.

[Go Top]

Q: Why on windows sometimes grid.py fails?

This problem shouldn't happen after version 2.85. If you are using earlier versions, please download the latest one.

[Go Top]

Q: Why grid.py/easy.py sometimes generates the following warning message?

Warning: empty z range [62.5:62.5], adjusting to [61.875:63.125]
Notice: cannot contour non grid data!

Nothing is wrong and please disregard the message. It is from gnuplot when drawing the contour.

[Go Top]

Q: How do I choose the kernel?

In general we suggest you to try the RBF kernel first. A recent result by Keerthi and Lin ( download paper here) shows that if RBF is used with model selection, then there is no need to consider the linear kernel. The kernel matrix using sigmoid may not be positive definite and in general it's accuracy is not better than RBF. (see the paper by Lin and Lin ( download paper here). Polynomial kernels are ok but if a high degree is used, numerical difficulties tend to happen (thinking about dth power of (<1) goes to 0 and (>1) goes to infinity).

[Go Top]

Q: How does LIBSVM perform parameter selection for multi-class problems?

LIBSVM implements "one-against-one" multi-class method, so there are k(k-1)/2 binary models, where k is the number of classes.

We can consider two ways to conduct parameter selection.

For any two classes of data, a parameter selection procedure is conducted. Finally, each decision function has its own optimal parameters.
The same parameters are used for all k(k-1)/2 binary classification problems. We select parameters that achieve the highest overall performance.

Each has its own advantages. A single parameter set may not be uniformly good for all k(k-1)/2 decision functions. However, as the overall accuracy is the final consideration, one parameter set for one decision function may lead to over-fitting. In the paper

Chen, Lin, and Schölkopf, A tutorial on nu-support vector machines. Applied Stochastic Models in Business and Industry, 21(2005), 111-136,

they have experimentally shown that the two methods give similar performance. Therefore, currently the parameter selection in LIBSVM takes the second approach by considering the same parameters for all k(k-1)/2 models.

[Go Top]

Q: How do I choose parameters for one-class SVM as training data are in only one class?

You have pre-specified true positive rate in mind and then search for parameters which achieve similar cross-validation accuracy.

[Go Top]

Q: Instead of grid.py, what if I would like to conduct parameter selection using other programmin languages?

For MATLAB, please see another question in FAQ.

For using shell scripts, please check the code written by Bjarte Johansen

[Go Top]

Q: Why training a probability model (i.e., -b 1) takes a longer time?

To construct this probability model, we internally conduct a cross validation, which is more time consuming than a regular training. Hence, in general you do parameter selection first without -b 1. You only use -b 1 when good parameters have been selected. In other words, you avoid using -b 1 and -v together.

[Go Top]

Q: Why using the -b option does not give me better accuracy?

There is absolutely no reason the probability outputs guarantee you better accuracy. The main purpose of this option is to provide you the probability estimates, but not to boost prediction accuracy. From our experience, after proper parameter selections, in general with and without -b have similar accuracy. Occasionally there are some differences. It is not recommended to compare the two under just a fixed parameter set as more differences will be observed.

[Go Top]

Q: Why using svm-predict -b 0 and -b 1 gives different accuracy values?

Let's just consider two-class classification here. After probability information is obtained in training, we do not have

prob > = 0.5 if and only if decision value >= 0.

So predictions may be different with -b 0 and 1.

[Go Top]

Q: How can I save images drawn by svm-toy?

For Microsoft windows, first press the "print screen" key on the keyboard. Open "Microsoft Paint" (included in Windows) and press "ctrl-v." Then you can clip the part of picture which you want. For X windows, you can use the program "xv" or "import" to grab the picture of the svm-toy window.

[Go Top]

Q: I press the "load" button to load data points but why svm-toy does not draw them ?

The program svm-toy assumes both attributes (i.e. x-axis and y-axis values) are in (0,1). Hence you want to scale your data to between a small positive number and a number less than but very close to 1. Moreover, class labels must be 1, 2, or 3 (not 1.0, 2.0 or anything else).

[Go Top]

Q: I would like svm-toy to handle more than three classes of data, what should I do ?

Taking windows/svm-toy.cpp as an example, you need to modify it and the difference from the original file is as the following: (for five classes of data)

30,32c30
< 	RGB(200,0,200),
< 	RGB(0,160,0),
< 	RGB(160,0,0)
---
> 	RGB(200,0,200)
39c37
< HBRUSH brush1, brush2, brush3, brush4, brush5;
---
> HBRUSH brush1, brush2, brush3;
113,114d110
< 	brush4 = CreateSolidBrush(colors[7]);
< 	brush5 = CreateSolidBrush(colors[8]);
155,157c151
< 	else if(v==3) return brush3;
< 	else if(v==4) return brush4;
< 	else return brush5;
---
> 	else return brush3;
325d318
< 	  int colornum = 5;
327c320
< 		svm_node *x_space = new svm_node[colornum * prob.l];
---
> 		svm_node *x_space = new svm_node[3 * prob.l];
333,338c326,331
< 			x_space[colornum * i].index = 1;
< 			x_space[colornum * i].value = q->x;
< 			x_space[colornum * i + 1].index = 2;
< 			x_space[colornum * i + 1].value = q->y;
< 			x_space[colornum * i + 2].index = -1;
< 			prob.x[i] = &x_space[colornum * i];
---
> 			x_space[3 * i].index = 1;
> 			x_space[3 * i].value = q->x;
> 			x_space[3 * i + 1].index = 2;
> 			x_space[3 * i + 1].value = q->y;
> 			x_space[3 * i + 2].index = -1;
> 			prob.x[i] = &x_space[3 * i];
397c390
< 				if(current_value > 5) current_value = 1;
---
> 				if(current_value > 3) current_value = 1;

[Go Top]

Q: What is the difference between Java version and C++ version of libsvm?

They are the same thing. We just rewrote the C++ code in Java.

[Go Top]

Q: Is the Java version significantly slower than the C++ version?

This depends on the VM you used. We have seen good VM which leads the Java version to be quite competitive with the C++ code. (though still slower)

[Go Top]

Q: While training I get the following error message: java.lang.OutOfMemoryError. What is wrong?

You should try to increase the maximum Java heap size. For example,

java -Xmx2048m -classpath libsvm.jar svm_train ...

sets the maximum heap size to 2048M.

[Go Top]

Q: Why you have the main source file svm.m4 and then transform it to svm.java?

Unlike C, Java does not have a preprocessor built-in. However, we need some macros (see first 3 lines of svm.m4).

[Go Top]

Q: Except the python-C++ interface provided, could I use Jython to call libsvm ?

Yes, here are some examples:

$ export CLASSPATH=$CLASSPATH:~/libsvm-2.91/java/libsvm.jar
$ ./jython
Jython 2.1a3 on java1.3.0 (JIT: jitc)
Type "copyright", "credits" or "license" for more information.
>>> from libsvm import *
>>> dir()
['__doc__', '__name__', 'svm', 'svm_model', 'svm_node', 'svm_parameter',
'svm_problem']
>>> x1 = [svm_node(index=1,value=1)]
>>> x2 = [svm_node(index=1,value=-1)]
>>> param = svm_parameter(svm_type=0,kernel_type=2,gamma=1,cache_size=40,eps=0.001,C=1,nr_weight=0,shrinking=1)
>>> prob = svm_problem(l=2,y=[1,-1],x=[x1,x2])
>>> model = svm.svm_train(prob,param)
*
optimization finished, #iter = 1
nu = 1.0
obj = -1.018315639346838, rho = 0.0
nSV = 2, nBSV = 2
Total nSV = 2
>>> svm.svm_predict(model,x1)
1.0
>>> svm.svm_predict(model,x2)
-1.0
>>> svm.svm_save_model("test.model",model)

[Go Top]

Q: I compile the MATLAB interface without problem, but why errors occur while running it?

Your compiler version may not be supported/compatible for MATLAB. Please check this MATLAB page first and then specify the version number. For example, if g++ X.Y is supported, replace

CXX = g++

in the Makefile with

CXX = g++-X.Y

[Go Top]

Q: On 64bit Windows I compile the MATLAB interface without problem, but why errors occur while running it?

Please make sure that you use the -largeArrayDims option in make.m. For example,

mex -largeArrayDims -O -c svm.cpp

Moreover, if you use Microsoft Visual Studio, probabally it is not properly installed. See the explanation here.

[Go Top]

Q: Does the MATLAB interface provide a function to do scaling?

It is extremely easy to do scaling under MATLAB. The following one-line code scale each feature to the range of [0,1]:

(data - repmat(min(data,[],1),size(data,1),1))*spdiags(1./(max(data,[],1)-min(data,[],1))',0,size(data,2),size(data,2))

[Go Top]

Q: How could I use MATLAB interface for parameter selection?

One can do this by a simple loop. See the following example:

bestcv = 0;
for log2c = -1:3,
  for log2g = -4:1,
    cmd = ['-v 5 -c ', num2str(2^log2c), ' -g ', num2str(2^log2g)];
    cv = svmtrain(heart_scale_label, heart_scale_inst, cmd);
    if (cv >= bestcv),
      bestcv = cv; bestc = 2^log2c; bestg = 2^log2g;
    end
    fprintf('%g %g %g (best c=%g, g=%g, rate=%g)\n', log2c, log2g, cv, bestc, bestg, bestcv);
  end
end

You may adjust the parameter range in the above loops.

[Go Top]

Q: I use MATLAB parallel programming toolbox on a multi-core environment for parameter selection. Why the program is even slower?

Fabrizio Lacalandra of University of Pisa reported this issue. It seems the problem is caused by the screen output. If you disable the info function using

#if 0,

then the problem may be solved.

[Go Top]

Q: How to use LIBSVM with OpenMP under MATLAB/Octave?

First, you must modify svm.cpp. Check the following faq, How can I use OpenMP to parallelize LIBSVM on a multicore/shared-memory computer?

To build the MATLAB/Octave interface, we recommend using make.m. You must append '-fopenmp' to CXXFLAGS and add '-lgomp' to mex options in make.m. See details below.

For MATLAB users, the modified code is:

mex CFLAGS="\$CFLAGS -std=c99" CXXFLAGS="\$CXXFLAGS -fopenmp" -largeArrayDims -I.. -lgomp svmtrain.c ../svm.cpp svm_model_matlab.c
mex CFLAGS="\$CFLAGS -std=c99" CXXFLAGS="\$CXXFLAGS -fopenmp" -largeArrayDims -I.. -lgomp svmpredict.c ../svm.cpp svm_model_matlab.c

For Octave users, the modified code is:

setenv('CXXFLAGS', '-fopenmp') 
mex -I.. -lgomp svmtrain.c ../svm.cpp svm_model_matlab.c
mex -I.. -lgomp svmpredict.c ../svm.cpp svm_model_matlab.c

If make.m fails under matlab and you use Makefile to compile the codes, you must modify two files:

You must append '-fopenmp' to CFLAGS in ../Makefile for C/C++ codes:

CFLAGS = -Wall -Wconversion -O3 -fPIC -fopenmp -I$(MATLABDIR)/extern/include -I..

and add '-lgomp' to MEX_OPTION in Makefile for the matlab/octave interface:

MEX_OPTION += -lgomp

To run the code, you must specify the number of threads. For example, before executing matlab/octave, you run

> export OMP_NUM_THREADS=8
> matlab

Here we assume Bash is used. Unfortunately, we do not know yet how to specify the number of threads within MATLAB/Octave. Our experiments show that

>> setenv('OMP_NUM_THREADS', '8');

does not work. Please contact us if you see how to solve this problem. On the other hand, you can specify the number of threads in the source code (thanks to comments from Ricardo Santiago-mozos):

#pragma omp parallel  for private(i) num_threads(8)

[Go Top]

Q: How could I generate the primal variable w of linear SVM?

Let's start from the binary class and assume you have two labels -1 and +1. After obtaining the model from calling svmtrain, do the following to have w and b:

w = model.SVs' * model.sv_coef;
b = -model.rho;

if model.Label(1) == -1
  w = -w;
  b = -b;
end

If you do regression or one-class SVM, then the if statement is not needed.

For multi-class SVM, we illustrate the setting in the following example of running the iris data, which have 3 classes

  
> [y, x] = libsvmread('../../htdocs/libsvmtools/datasets/multiclass/iris.scale');
> m = svmtrain(y, x, '-t 0')

m = 

    Parameters: [5x1 double]
      nr_class: 3
       totalSV: 42
           rho: [3x1 double]
         Label: [3x1 double]
         ProbA: []
         ProbB: []
           nSV: [3x1 double]
       sv_coef: [42x2 double]
           SVs: [42x4 double]

sv_coef is like:

+-+-+--------------------+
|1|1|                    |
|v|v|  SVs from class 1  |
|2|3|                    |
+-+-+--------------------+
|1|2|                    |
|v|v|  SVs from class 2  |
|2|3|                    |
+-+-+--------------------+
|1|2|                    |
|v|v|  SVs from class 3  |
|3|3|                    |
+-+-+--------------------+

so we need to see nSV of each classes.

  
> m.nSV

ans =

     3
    21
    18

Suppose the goal is to find the vector w of classes 1 vs 3. Then y_i alpha_i of training 1 vs 3 are

  
> coef = [m.sv_coef(1:3,2); m.sv_coef(25:42,1)];

and SVs are:

  
> SVs = [m.SVs(1:3,:); m.SVs(25:42,:)];

Hence, w is

> w = SVs'*coef;

For rho,

> m.rho

ans =

    1.1465
    0.3682
   -1.9969
> b = -m.rho(2);

because rho is arranged by 1vs2 1vs3 2vs3.

[Go Top]

Q: Is there an OCTAVE interface for libsvm?

Yes, after libsvm 2.86, the matlab interface works on OCTAVE as well. Please use make.m by typing

>> make

under OCTAVE.

[Go Top]

Q: How to handle the name conflict between svmtrain in the libsvm matlab interface and that in MATLAB bioinformatics toolbox?

The easiest way is to rename the svmtrain binary file (e.g., svmtrain.mexw32 on 32-bit windows) to a different name (e.g., svmtrain2.mexw32).

[Go Top]

Q: On Windows I got an error message "Invalid MEX-file: Specific module not found" when running the pre-built MATLAB interface in the windows sub-directory. What should I do?

The error usually happens when there are missing runtime components such as MSVCR100.dll on your Windows platform. You can use tools such as Dependency Walker to find missing library files.

For example, if the pre-built MEX files are compiled by Visual C++ 2010, you must have installed Microsoft Visual C++ Redistributable Package 2010 (vcredist_x86.exe). You can easily find the freely available file from Microsoft's web site.

For 64bit Windows, the situation is similar. If the pre-built files are by Visual C++ 2008, then you must have Microsoft Visual C++ Redistributable Package 2008 (vcredist_x64.exe).

[Go Top]

Q: LIBSVM supports 1-vs-1 multi-class classification. If instead I would like to use 1-vs-rest, how to implement it using MATLAB interface?

Please use code in the following directory. The following example shows how to train and test the problem dna (training and testing).

Load, train and predict data:

[trainY trainX] = libsvmread('./dna.scale');
[testY testX] = libsvmread('./dna.scale.t');
model = ovrtrain(trainY, trainX, '-c 8 -g 4');
[pred ac decv] = ovrpredict(testY, testX, model);
fprintf('Accuracy = %g%%\n', ac * 100);

Conduct CV on a grid of parameters

bestcv = 0;
for log2c = -1:2:3,
  for log2g = -4:2:1,
    cmd = ['-q -c ', num2str(2^log2c), ' -g ', num2str(2^log2g)];
    cv = get_cv_ac(trainY, trainX, cmd, 3);
    if (cv >= bestcv),
      bestcv = cv; bestc = 2^log2c; bestg = 2^log2g;
    end
    fprintf('%g %g %g (best c=%g, g=%g, rate=%g)\n', log2c, log2g, cv, bestc, bestg, bestcv);
  end
end

[Go Top]

Q: I tried to install matlab interface on mac, but failed. What should I do?

We assume that in a matlab command window you change directory to libsvm/matlab and type

>> make

We discuss the following situations.

An error message like "libsvmread.c:1:19: fatal error: stdio.h: No such file or directory" appears.
Reason: "make" looks for a C++ compiler, but no compiler is found. To get one, you can
- Install XCode offered by Apple Inc.
- Install XCode Command Line Tools.
On OS X with Xcode 4.2+, I got an error message like "llvm-gcc-4.2: command not found."
Reason: Since Apple Inc. only ships llsvm-gcc instead of gcc-4.2, llvm-gcc-4.2 cannot be found.

If you are using Xcode 4.2-4.6, a related solution is offered at http://www.mathworks.com/matlabcentral/answers/94092.

On the other hand, for Xcode 5 (including Xcode 4.2-4.6), in a Matlab command window, enter
- cd (matlabroot)
- cd bin
- Backup your mexopts.sh first
- edit mexopts.sh
- Scroll down to "maci64" section. Change
```
		CC='llvm-gcc-4.2'
		CXX='llvm-g++-4.2'
```
  to
```
		CC='llvm-gcc'
		CXX='llvm-g++'
```
Please also ensure that SDKROOT corresponds to the SDK version you are using.
Other errors: you may check http://www.mathworks.com/matlabcentral/answers/94092.

[Go Top]

Q: I tried to install octave interface on windows, but failed. What should I do?

This may be due to that Octave's math.h file does not refer to the correct location of Visual Studio's math.h. Please see this nice page for detailed instructions.

[Go Top]

LIBSVM home page

我研究了好久才发现有这么个好东西的:转别人的

问: 我在那里能够找到libsvm的文件 ?

软件包中有一个 README 文件，里面详细说明了所有参数选项、数据格式以及库函数的调用。在python目录下，模型选择工具和python界面的libsvm各有一个README文件。初学者可以通过A practical guide to support vector classification 了解如何训练和检验数据.论文LIBSVM : a library for support vector machines详细讨论了libsvm的使用.

[Go Top]

问: 以前版本的libsvm都有什么变化?
详见变化日志. 你可以到这里下载以前版本的libsvm .

[Go Top]

问: 如果我想引用libsvm，我应该引用哪篇论文?
请引用以下论文: Chih-Chung Chang and Chih-Jen Lin, LIBSVM : a library for support vector machines, 2001. Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm bibtex格式如下：

@Manual{CC01a,
author =	 {Chih-Chung Chang and Chih-Jen Lin},
title =	 {{LIBSVM}: a library for support vector machines},
year =	 {2001},
note =	 {Software available at \url{http://www.csie.ntu.edu.tw/~cjlin/libsvm}}
}

[Go Top]

问: 如果我想在我的软件中应用svm，是否存在软件许可问题？
Libsvm许可文件 ("修改了的BSD许可文件") 和其他许多软件的许可文件都是兼容的，如 GPL. 因此你很容易就可以把libsvm应用到你的软件中，你还可以将它应用到商业产品中.

[Go Top]

问: 是否有一系列基于libsvm的附加工具?
是的，详见libsvm tools

[Go Top]

问: 在Unix操作系统中，出现了"error in loading shared libraries"（“在装载共享库时发生错误”）或者"cannot open shared object file."（无法打开共享目标文件“）这是为什么?
如果你在一台机器上编译代码，然后到另外一台有着不兼容库的机器上运行，这种情况酒会经常发生。你在那台机器上可以重新编译一下，或者用静态连接。

[Go Top]

问:修改源代码后，如果想在MS窗口中建立一个“svm-toy”的图形界面，我该怎么做？
你可以通过选择 "Win32 Project"来建立一个工程。另外对于"svm-train"和 "svm-predict" ，你需要选择"Win32 Console Project"。对于libsvm2.5以后的版本，你还可以通过Makefile.win来实现。详见README。

如果你没有用Makefile.win并出现了以下连接错误：LIBCMTD.lib(wwincrt0.obj) : error LNK2001: unresolved external symbol
_wWinMain@16
那么，你很有可能是选择了一个错误的工程类型。

[Go Top]

问: 我是MS Windows用户，在这些预编译的exe文件中，为什么只有一个(SVM_toy)运行? 你应当打开命令窗口，然后输入svmtrain.exe，这样就可以在DOS窗口中看到所有的选项了。你可以到README文件中看到一些例子

[Go Top]

问:为什么有时我在training/model 文件中看不到所有的数据属性呢？W
libsvm应用了所谓的”稀疏“格式，这样零值就不用存储了。例如，有下面属性的数据

1 0 2 0
将被替换为： 1:1 3:2

[Go Top]

问:如果我的数据是非数值型的，可以用libsvm吗？目前libsvm只支持数值型的数据。因此，你必须将非数值型的转为数值型的数据。比如，你可以用二进制属性来替代原来的类别属性。

[Go Top]

问:为什么要采用稀疏格式呢? 密集数据在训练时候会不会很慢？这是个具有争议的话题。将系数向量赋值给核函数是比较慢的，因此总的训练时间至少是采用密集格式的2倍或3倍。但是，我们不支持密集格式的数据，因为我们不能够处理极度稀疏的数据。代码的简洁也是我们考虑的一个因素。目前我们决定只支持稀疏格式的数据。

[Go Top]

问: 训练C-SVM时，其输出结果如下，它们代表什么意思?
optimization finished, #iter = 219
nu = 0.431030
obj = -100.877286, rho = 0.424632
nSV = 132, nBSV = 107
Total nSV = 132 obj 是对偶SVM问题的最优目标值. rho 是判决函数sgn(w^Tx - rho)的偏项. nSV 和 nBSV 分别代表支持向量和边界支持向量 (即alpha_i = C)的个数. nu-svm 在某中程度上可以看作 C-SVM的等价形式，不过其中的 C被替换为 nu. nu 仅表明了相应的参数. 更详细的解释请看libsvm document.

[Go Top]

问:关于模型文件，您能否解释的更详细一点？
在参数后面，每一条线代表一个支持向量。支持向量按照先列出的标签列出（即类标签列表中，第一类的支持向量首先分组，然后依次类推）

如果k是类别的总数，那么每个支持向量的前面就有k-1个y*alpha系数，其中alpha代表如下二分类问题的对偶解：
1 vs j, 2 vs j, ..., j-1 vs j, j vs j+1, j vs j+2, ..., j vs k
并且在前j-1个系数中，y=+1, 其余的k-j个系数中y=-1.例如，对于4分类问题，文件看起来结构如下：

+-+-+-+--------------------+
|1|1|1|                    |
|v|v|v|  类别1的支持向量  |
|2|3|4|                    |
+-+-+-+--------------------+
|1|2|2|                    |
|v|v|v|  类别2的支持向量 |
|2|3|4|                    |
+-+-+-+--------------------+
|1|2|3|                    |
|v|v|v|  类别3的支持向量  |
|3|3|4|                    |
+-+-+-+--------------------+
|1|2|3|                    |
|v|v|v|  类别4的支持向量  |
|4|4|4|                    |
+-+-+-+--------------------+

问: 在缓缓存中，是否可以用浮点型或者双精度型来存储数据? 
当你在缓存中存入更多的数据时，我们默认的是浮点型.一般情况下，这样就很好了，但是对于少数困难的情况（如C非常非常大）解将是巨大的数字。这时，仅用浮点型，数值的精度可能就不够了。

[Go Top]

问：怎样选择核函数?

通常我们建议你首先采用RBF核函数。Keerthi 和 Lin 的最近的研究( 下载论文) 表明如果模型选择RBF的话，就没有必要再考虑线性核函数了。采用sigmoid核函数的矩阵不一定会正有界，而且通常它的准确性也不如RBF(可参见Lin和Lin的论文此处下载). 多项式核函数还不错，但是如果度数很高的话，数值困难就会发生(考虑到(<1)的d次幂会趋向0，(>1)的d次幂会趋向无穷）

[Go Top]

问: libsvm是否可以用来专门处理线性SVM？D

不是的，目前libsvm用同样的方法处理线性/非线性SVMs. 注意：如果采用线性核函数，一些技巧可能会节省训练/检验的时间。因此libsvm对线性SVM并不时特别有效,尤其是采用很大的C的问题，这些问题数据的数值比其属性值要大得多。你可以：

仅用很大的C.下面的论文表明了：当C大于一个确定的阀值以后，判决函数是相同的。
S. S. Keerthi and C.-J. Lin. Asymptotic behaviors of support vector machines with Gaussian kernel . Neural Computation, 15(2003), 1667-1689.
尝试用bsvm, 它有个对解决线性SVMs很有效的方法.你可以在下面的研究中找到更详细的内容:
K.-M. Chung, W.-C. Kao, T. Sun, and C.-J. Lin. Decomposition Methods for Linear Support Vector Machines. Neural Computation, 16(2004), 1689-1704.

另外，你并没必要一定要解决线性SVMs.你可以参考前面有关如何选取核函数的问题。

[Go Top]

问: 自由支持向量特别大，我该怎么办?

当数据过拟和时，这种情况会经常发生. 如果数据的属性值范围特别大，你可以尝试调整它们的范围。这样合适参数的域可能会很大。注意：libsvm里包含有数据调整程序的，即svm_scale.

[Go Top]

问: 我是否可以用同样的方法来调整训练数据和检验数据的范围？

是的, 可以按照以下方法:
svm-scale -s scaling_parameters train_data > scaled_train_data
svm-scale -r scaling_parameters test_data > scaled_test_data

[Go Top]

问:将属性值限制到[0,1]，是否比限制到 [-1,1]有很大的不同?

对于线性规划方法，如果采用可RBF核函数并进行了参数的选择，两者是没什么不同的。假设Mi和mi分别代表第i个属性的最大值和最小值.规划到[0,1]即：

                x'=(x-mi)/(Mi-mi)

对于[-1 1]：

                x''=2(x-mi)/(Mi-mi)-1.

在RBF核函数中：

                x'-y'=(x-y)/(Mi-mi), x''-y''=2(x-y)/(Mi-mi).

因此在[0,1]数据规划中用(C,g)，和[-1 1]数据规划中用(C,g/2)是一样的。

尽管性能相同，但是计算机时间可能不同。对于有许多零入口的数据, [0,1]规划保持了输入数据的稀疏特性，因此可能会节省计算时间。

[Go Top]

问: 预测准确率很低，应该如何改进？

可以用python目下的grid.py来寻找合适的参数，grid.py是一个用来选择模型参数的工具。关于模型选择的重要性，你可以参考我的谈话： A practical guide to support vector classification

[Go Top]

问: 我的数据是不平衡的，libsvm能解决这样的问题吗？

可以。libsvm有一个-wi选项。例如，你用：

svm-train -s 0 -c 10 -w1 1 -w-1 5 data_file

则对类别“-1”的惩罚就较大。注意-w选项仅用在C-SVC中。

[Go Top]

问: nu-SVC和C-SVC有什么不同之处?
除了参数不同外，两者基本是一样的。C-SVC中，C的范围是0到无穷，nu-SVC中C的范围是[0 1]。 nu一个很好的特性是：它与支持向量的比率和训练误差的比率相关。

[Go Top]

问: 程序一直运行，但却没有输出，应该怎么办？

你最好检查一下数据. 每一个训练/学习的数据都必须在一行，不可以分开。此外，你必须删除空行。

[Go Top]

问: 程序一直运行（有输出，比如许多点）。应该怎么办？

从理论上讲，如果核矩阵式半正定的，libsvm可以保证收敛。2.4以后版本的libsvm还可以处理非半正定矩阵核，比如sigmoid(tanh)核。这样看来，你处理的情况是病态情况（比如参数过大或过小），这样机会发生数值困难。

[Go Top]

问: 训练时间太长，应该怎么办？

对于大型问题，请确保有足够的缓存空间（即-m）。对于一些困难的情况（比如-c很大），会出现收敛很慢的情况. 你可以常识用一个较大点的终止可容忍偏差. 如果这样仍然不行，你可以和我们联系。我们会告诉你一些关于改善训练时间短小技巧。

[Go Top]

问:我如何才能得到判决值?

对于回归，程序可以打印出判决值。对于分类，可以通过解几个二进制SVMs来实现多分类，因此你可以通过调用子程序svm_predict_values来实现输出判决值。然后可以通过子程序svm_get_labels得到相应的标签值。详见软件包中的README文件。

我们并不推荐以下操作，但是如果你要得到二分类的标签值+1和-1（注：这里的+1、-1和5、10的含义是不一样的），你只要在svm.cpp文件中的 svm_predict_values(model, x, dec_values) 后面添加：printf("%f\n",dec_values[0]*model->label[0])就可以了。

正（负）判决值对应的值为+1（-1）

[Go Top]

问:怎样得到一个点到超平面的距离？

距离的表达式为： |decision_value| / |w|.我们有： |w|^2 = w^Tw = alpha^T Q alpha = 2*(dual_obj + sum alpha_i). 因此在svm.pp文件中，知道输出对偶目标值的语句，在后面加上打印输出 w^Tw的语句就可以了.

[Go Top]

问:在linux操作系统上，对于一些问题，如果用很大的缓存（即很大的-m），有时候会出现“段错误”（"segmentation fault“），为什么?

在32位机器上，最大可设地址空间为4GB。Linux核将其分成3：1，即：用户的空间为3GB，核空间为1G。尽管用户有3GB的存储空间，但是最大动态分配内存只有2GB。因此如果你将-m设为2G，内存将被耗尽。因此svm-train运行时候如果需要更多内存，程序就无法运行。详见此文.

有两中方法可以解决这个问题。如果你的机器支持Intel的物理地址扩展，你可以在Linux核中起用HIGHMEM64G选项，这样用户和核都将分配到4G的空间. 此外, 你还可以使用`tub'软件，它将为动态分配内存 elimate the 2G boundary . `tub' 可以到http://www.bitwagon.com/tub.html下载.

[Go Top]

问：: 怎样才可以不让svm-train 和svm-predict 在屏幕上输出?

只要将svm.cpp中的:

#if 1
void info(char *fmt,...)

改为：

#if 0
void info(char *fmt,...)

[Go Top]

问:我想用自己的核函数，但我发现两个核函数赋值子程序 k_function()和kernel_function().我应该修改哪个？

[Go Top]

问:对于多分类SVM，libsvm采用的是什么方法 ? 为什么不用"1-against-the rest" 法 ?

对于多分类，我们采用的是1against 1法.我们的选择建立在以下对比的基础上: C.-W. Hsu and C.-J. Lin. A comparison of methods for multi-class support vector machines , IEEE Transactions on Neural Networks, 13(2002), 415-425.

"1agains1the rest"是个很好的方法，而且分类效果和"1-against-1."可以相比。但是我们采用后者，因为它训练的时间更短。

[Go Top]

:做完交叉检验以后，为什么没有模型文件输出?

交叉检验只是为了寻找好的参数，寻找完后，你必须在不用-v选项的情况下，重新训练整个数据，才会得到模型文件。

[Go Top]

问:对于交叉检验，我想使用随机划分，应该如何操作？

如果你用的是 GNU C库，默认的种子是1.因此你在运行svm-train -v后，总是得到相同的结果. 想用不同的种子，你可在svm-train.c中添加如下代码：

#include <time.h>

在子程序 do_cross_validation()的开始部分添加,

srand(time(0));

[Go Top]

问:如果我想解决L2-svm 问题(即二次方误差项). 我应该怎样修改代码 ?

这十分简单. 以c-svc为例, 在svm.cpp中只要修改两个地方即可.

第一, 将solve_c_svc中的：

	s.Solve(l, SVC_Q(*prob,*param,y), minus_ones, y,
alpha, Cp, Cn, param->eps, si, param->shrinking);

修改为：

	s.Solve(l, SVC_Q(*prob,*param,y), minus_ones, y,
alpha, INF, INF, param->eps, si, param->shrinking);

第二：在SVC_Q类中, 声明C为私有变量：

	double C;

在构造（constructor）中，将它赋给param.C:

        this->C = param.C;

在子程序get_Q中, 在for循环之后，添加：

        if(i >= start && i < len)
data[i] += 1/C;

对于一分类svm，以上修改完全一样。对于SVR,上面的if语句就没有必要了，你只要用一个简单的赋值语句即可：

	data[real_i] += 1/C;

LIBSVM的FQA

LIBSVM FAQ

猜你喜欢