Welcome to LabVIEW Machine Learning Toolkit

[DSPGuy]

Fuzzy Logic is covered in a different toolkit, and is oriented more to control applications.  The Machine Learning toolkit has a neural network implementation and validation tools.  Even if you write your own neural network implementation, the valalidation functions are nice to have.

[duongducthien]

Is there any one makes Mel Frequency Ceptrum Coefficients (MFCCs) in this toolkit? It should be have in this one, doesn't it?

[James_Covington]

Hi,

Is there a 64bit version of MLT? If so, please could you give me the link. Many thanks.

[duetcat]

MFCC is not included in this toolkit currently. However, all the 'parts' needed by this algorithm (FFT, DCT, ect) are avaiable in LabVIEW core. You could easily implement this algorithm within LabVIEW.

[duetcat]

MLT is purely written in G, so it can work with 64bit LV directly. If you already installed MLT in 32-bit LV, just simply copy

C:\Program Files (x86)\National Instruments\LabVIEW 2012\vi.lib\addons\Machine Learning

to

C:\Program Files\National Instruments\LabVIEW 2012\vi.lib\addons\Machine Learning

Or you can visit the homepage

https://decibel.ni.com/content/docs/DOC-19328

download the zip file: Machine Learning Toolkit Manually Install.zip

and extract it to

C:\Program Files\National Instruments\LabVIEW 2012\vi.lib\addons\Machine Learning

[Vysh]

Hi,

I am tring to clssify audio into two classes. I have 11 samples and 9 features for each sample. My training data has a dimension of 11*9. I am not very sure about how to give the input to the k-nn classifier. Plus, I am not very sure about the distance input that needs to be given. Can you tell me how to use the k-nn classifier that comes with the machine learning toolkit?

Thank you,

Vysh

[zety]

hi, can we use Labview for SIFT descriptor?

[muku64]

please can you suggest me,From where the detailed help for ' machine learning toolkit icons' can be downloaded?

[dork_matter]

I have applied the shipped MLT example: "Example_BP Network_Classification.vi" succesfully with two of my own generated training and test data sets.  However, I have generated recent data sets that result in a Training MSE value of "NaN."  I have verified that the both the training set data and training set classes are in the correct format.  I can't think of any reason why these subsequent training data sets error out.  Could you please provide guidance on troubelshooting this error?

[joesze]

Some combinations of data and number of hidden neurons will fail to

converge. Try using some other numbers of hidden neurons and see if you can

find a combination that converges for your data.

Joe