MSE Vs Training Cycle in Neural Network
MF_Hussain
New Altair Community Member
Hello,
Hope this message finds the reader in good health. I am new to Neural Network in general and RapidMiner in particular. Using the Supervised Learning (Neural Net) to train on my data set. I wish to use the NN Model and specify 5000 cycles to it but wish to have an MSE vs Cycles plot at the end to see at which cycle the error reaches close to my acceptable levels. Kindly help as in what to do to achieve this task.
Regards
p.s. I am using RapidMider 9.7
Hope this message finds the reader in good health. I am new to Neural Network in general and RapidMiner in particular. Using the Supervised Learning (Neural Net) to train on my data set. I wish to use the NN Model and specify 5000 cycles to it but wish to have an MSE vs Cycles plot at the end to see at which cycle the error reaches close to my acceptable levels. Kindly help as in what to do to achieve this task.
Regards
p.s. I am using RapidMider 9.7
Tagged:
0
Best Answer
-
Hi @MF_Hussain
Welcome to the community.
When you want to go deeper with using Deep Learning in RapidMiner it is worthwhile to install the Deep Learning extension (install it over the marketplace) and use the operators provided by the extension to configure your neural networks. While the bundled Neural Net and Deep Learning operators are built-in implementations which should be a bit easier to use, the Deep Learning extension is way more powerful in using Deep Learning in RapidMiner. For example it provides you with the history port which allows exactly what you are looking for. The visualization of the error vs the trained cycles (or epochs).
You can use this presentation: https://rapidminer.com/resource/deep-learning/ to learn more about Deep Learning and RapidMiner in general.
Hopes this helps
Best regards,
Fabian1
Answers
-
Hi @MF_Hussain
Welcome to the community.
When you want to go deeper with using Deep Learning in RapidMiner it is worthwhile to install the Deep Learning extension (install it over the marketplace) and use the operators provided by the extension to configure your neural networks. While the bundled Neural Net and Deep Learning operators are built-in implementations which should be a bit easier to use, the Deep Learning extension is way more powerful in using Deep Learning in RapidMiner. For example it provides you with the history port which allows exactly what you are looking for. The visualization of the error vs the trained cycles (or epochs).
You can use this presentation: https://rapidminer.com/resource/deep-learning/ to learn more about Deep Learning and RapidMiner in general.
Hopes this helps
Best regards,
Fabian1