Question

In MATLAB neural net fitting app why cant i select different input and target data? i select the input data then, when i select target data the input becomes the same file automatically? how can i solce the problem to train the network?

input data is 3831 samples of 4 elements

target is 1643 samples of 4 elements

IJ DE Intelliu IDEA Eclipse Java Visual Studio CodeBlocks RStudio Rx64 3.5.0 R i386 3.5.0 XAMPP Wampserve Oxygen yzer MATLAB Coder Application Classification DistributionNeural Net Neural Net Neural Net Neural Net ata and targets define your fitting problem? Summary pace to the network. inputs 'data is a 1643x4 matrio, representing static data: 1643 samples of 4 data ..elements esired network output. Targets data is a 1643x4 matrix, representing static data: 1643 samples of 4 elements. data ON:atrix columns ๏ Matrix rows ool with an example data set? Load Example Data Set dick [Next Start N4 Welcome 1. Back L-Net l L°Cancel Back Next

Answer 1

I have tried to create a random matrix according to size of your input and target data.

Please follow the screenshots and do the same to select the specific target data.

step 1: initiate matlab and load data sets and type "nntool" on command prompt

Command Window Current Folder Name Icdata.xsd Icdata_utf8.xml >>input-randi (0, 5], [3841,4]) >> output-randi (0,5], [1643,4]); Icdata.xml mexutils.pm mexsetup.pm mex.pl license standalone.lic matlab.exe worker.bat mw_mpiexec.bat mexext.bat mex.bat 囚mcc.bat mbuild.bat Details Workspac Name Value input 3841x4 double

step 2: click on import

Open VariableRun and Time Save orkspace Clear WorkspaceClear Commands Simulink Layout Neural Network/Data Manager (nntool) CODE Networks Input Data: Output Data: ATLAB R2016a bin Command Window >>input-randi ([0, 5], [3841,4]) >output-randi ([0, 5], [1643, 4]) > nntool Error Data: Target Data: Layer Delay States: Input Delay States: DHelpClose Import..New...Ope 其Delete

step 3: select variable and destination as import data

Code Analyze O Preferences New Variable Open vanable ▼ Clear Workspace 2 帛Import to Network/Data Manager Source ⓔ Import from MATLAB workspace Select a Variable Output Data . (n selectio Name input Import As input Load from disk file mand Window utput > input-randi (0 MAT-file Name output-randi nntool Network 0 Target Data O Initial Input States OInitial Layer States O Output Data O Error Data Error Data: Close Layer Delay States: Input Delay States; X Delete New. Import.. Help Close

step 4: again select variable for target data as destination

Code Analyze O Preferences New vanable Open Variable Import to NetwotkiData Manager Savg orkspaceClear Workspace Source Sclect a Variablc Destination Name output Import As -」Output Data: Import from MATLAB workspace OLoad from disk file (no selection) input Command Window >>input-randi (IO MAT-file Name >>output-randi Network >nntool O Input Data O Initial input States O Initial Layer States O Output Data Error Data: | 0 Error Data Close Layer D ッInput Delay States: States: Help O Close New Import...

step 5: see in input and output data there are entities and now click on "new".

New V Analyze CodeO Prelerance Praferences Community Opan Variableun and Tima Run and Time Save ▼u- Neural Network/Data Manager (nntool) /Clear Commands ▼ Simulink Layout 事Networks Input Data: t Data: MATLAB R2016a bin Command Window >> input-randi (t0,5], [3841,41) >> output randi ([0,51, [1643,41) >nntool Target Data: Error Data: States: States: Import... R Delete Export... Help Close

step 6: now again a pop up will appear and select input data.

New Variable Analyze Code Run and Time Clear Commands ▼ Open Vaiable Simulink Layout Save orkspaceClear Workspace Neural Network/Data Manager (nn tool) CODE , Input Data; input Create Network or Data ATLABR2016a bin Network Data Command Window >>input-randi (I0,51, [3841,41) >output randi ([0,5), (1643,4]) >> nntool Name network Network Properties Network Type: Feed-forward backprop Target Input data: Target data Training function: Adaption leaming function: Performance function: Number of layers: (Select a Target) TRAINLM LEARNGDM MSE Properties for Layer 1 States: t Dela Number of neurons: 10 TANSIG Transfer Function: View Restore Defaults ☆ Create Help Close Help O Close Import

step 7: now select for output data

tz Open Variable ▼ WorkspaceClear Workspace O Run and Time Simulink Save Layout!参Neural Network/Data Manager (nntool CODE , Input Data Create Network or Data MATLABR2016a bin Network Data O Command Window >>input-randi ([0,5), [3841,4) A>> output-randi ([0,51, [1643, 41) >nntool Name network Network Properties Network Type: Feed-forward backprop O Tar Input data: Target data: Training function: Adaption learning function: Performance function: Number of layers: input TRAINLM LEARNGDM MSE Properties for: Layer 1 t Dela States: Number of neurons: 10 Transfer Function: TANSIG View Restore Defaults er create * Close Help Help Close Import

step 8: now click on "create" to create the network.

@ Preferences New VaiableAnalyze Code Open Variable Run and Time Import Save Data Workpace Clear Workspace Neural Network/Data Manager (nntool) Input DataC input /Clear Commands ▼ Simulink Layout Clear Create Network or Data FilesMATLABR2016a bin Network Data Command Windovw >>input-randi (I0,51, [3841,41) Name >output-randi ([0, 5], [1643,4]) >nntool Network Properties Network Type: Feed-forward backprop Target Dat Input data: Target data: Training function: Adaption leaming function: Performance function: Number of layers: input output TRAINLNM LEARNGDM MSE for: Layer 1v Dela States: Number of neurons: 10 Transfer Function: TANSIG □View ☆ Restore Defaults Help Close Import Help Close