How can I create a custom neural network by this following data set?
1 回表示 (過去 30 日間)
古いコメントを表示
I have to take 4 inputs (Independent data array) and 2 inputs (dependent data array). I have 18 sets of data. The number of layers are 2.
Thanks
0 件のコメント
採用された回答
Greg Heath
2015 年 12 月 15 日
I see no reason to use a custom network.
Use fitnet for regression or curvefitting
Use patternnet for classification or pattern recognition.
Use the commands help and doc to access the documentation that includes an example for each function.
For example:
help fitnet
doc patternnet
Hope this helps.
Greg
2 件のコメント
Greg Heath
2015 年 12 月 16 日
In general, I recommend transforming inputs and outputs. Mapminmax is the default, i.e., it will be implemented automatically. If you wish, you can remove it or change to mapstd for zero-mean/unit-variance:
help mapstd
doc mapstd
However, I like to implement that transformation explicitly before training by using zscore:
help zscore
doc zscore
Then I can modify or delete outliers before training. Since I am lazy, I just accept the default mapminmax instead of removing it.
I have posted zillions of examples in both the NEWSGROUP and ANSWERS. Include greg and zscore in your search words.
Hope this helps.
Thank you for formally accepting my answer
Greg
その他の回答 (0 件)
参考
カテゴリ
Help Center および File Exchange で Sequence and Numeric Feature Data Workflows についてさらに検索
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!