In mat2py how to fix error AttributeError: 'numpy.ndarray' object has no attribute 'get_svm_type'

Question

I am trying to convert Matlab code to Python code. I have downloaded the libsvm library using sudo apt-get install python-libsvm but I am getting an error.

import scipy.io
from svmutil import svm_predict
def SVM_LIBSVM(VECTOR_TRAINING_LABELS , VECTOR_TRAINING_POINTS ,VECTOR_TESTING_LABELS , VECTOR_TESTING_POINTS):
    mat = scipy.io.loadmat('model.mat')
    model = mat['model']
    predicted_label = list()
    predicted_label = svm_predict(VECTOR_TESTING_LABELS , VECTOR_TESTING_POINTS , model)
    return predicted_label

I am getting the following error:

File "/home/optimum/mat2py/svmutil.py", line 193, in svm_predict
    svm_type = m.get_svm_type()
AttributeError: 'numpy.ndarray' object has no attribute 'get_svm_type'

peterfranciscook · Accepted Answer · 2019-05-08 15:53:51Z

1

I'm assuming that model object is from the MATLAB libsvm interface, which would make it a MATLAB struct and not a svm_model object. AFAIK there is not a direct conversion, but you should be able to read the fields from the struct, which you may be able to use for a python svm_model object (not 100% on this; I haven't used the python API much).

model_parameters = model['Parameters']
model_nr_class = model['nr_class']
model_total_sv = model['totalSV']
model_rho = model['rho']
model_label = model['Label']
model_sv_indices = model['sv_indices']
model_prob_a = model['ProbA']
model_prob_b = model['ProbB']
model_n_sv = model['nSV']
model_sv_coef = model['sv_coef']
model_svs = model['SVs']

The most painful way to generate this, if you can't find a way in python, would be to move the MATLAB struct values to a ctypes struct then use toPyModel (python API) to convert it to a python svm_model.

struct svm_model
{
    struct svm_parameter param; /* parameter */
    int nr_class;       /* number of classes, = 2 in regression/one class svm */
    int l;          /* total #SV */
    struct svm_node **SV;       /* SVs (SV[l]) */
    double **sv_coef;   /* coefficients for SVs in decision functions (sv_coef[k-1][l]) */
    double *rho;        /* constants in decision functions (rho[k*(k-1)/2]) */
    double *probA;      /* pairwise probability information */
    double *probB;
    int *sv_indices;        /* sv_indices[0,...,nSV-1] are values in [1,...,num_traning_data] to indicate SVs in the training set */

    /* for classification only */

    int *label;     /* label of each class (label[k]) */
    int *nSV;       /* number of SVs for each class (nSV[k]) */
                /* nSV[0] + nSV[1] + ... + nSV[k-1] = l */
    /* XXX */
    int free_sv;        /* 1 if svm_model is created by svm_load_model*/
                /* 0 if svm_model is created by svm_train */
};

answered May 8, 2019 at 15:53

peterfranciscook

3112 silver badges7 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Optimum Python Over a year ago

thnks ..can you give me link to the python api

Optimum Python Over a year ago

thanks for replying,can't we do it in pure python like using svm_predict function in package mat2py ? i used it but i am getting the above error i am using the following package github.com/edersoncorbari/mat2py

Collectives™ on Stack Overflow

In mat2py how to fix error AttributeError: 'numpy.ndarray' object has no attribute 'get_svm_type'

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related