I am looking for another modeling tools that is automated or programmed instead of ANN modeling, and is it possible to make exact and confidence outputs from those software ? Actually i am trying to use ANN and other related neural networking but it requires a lot of time from researcher and just want to know if there are any other simplified tools or programs
I would check tensorflow.org and https://github.com/mila-udem/blocks in my opinion those are best toolkits currently avaible
Hi Usukhbayar.
For unsupervised learning ANN is the best technique so far. But for simpler implementation you can use any of these techniques: Association rule learning, Hierarchical clustering, Cluster analysis, Outlier Detection
You can use free WEKA software for modeling.
visit the below website:
http://www.cs.waikato.ac.nz/ml/weka/
Hi,Usukhbayar! The BEST programm for the ANN-modiling is the MatLAB (Neural Network Toolbox). Also can recommende for you "Statistica NeuralNetworks".
In my experience, Matlab is the most powerful software avaialble for ANN. However, there are still limitations to Matlab toolbox. To have ultimate control over your neural network architecture, you need to develop your own code.
You can use free WEKA software for predictive purposes. See the below link:
http://www.cs.waikato.ac.nz/ml/weka/
more over in MatLAB you can not only realise ANN-modeling but can get executional program (exe-file) with your trained ANN. in this case you save many hours for writing and testing the program code in a high level language. the choice is yours))
crate in put , output and target data then open anntool using MATLAB.
Than train the network
NN toolbox in MATLAB is a nice option, but if you need much analysis without going into the nuances of MATLAB, 'Weka' tools is a very good option, moreover its a freeware and open source software.
Dear,
Arnab Sadhu ,
I agree with you, i have also used ANN tool built in matlab, Its user friendly and one can easily train the network to predict the new variables.
Nasar
- Badges
- Science topic