Deep Learning + SIMULINK + BLOK PREDICT = Regression
1 次查看(过去 30 天)
显示 更早的评论
How to provide data to the CNN machine in SIMULINK with a timestamp. I chose the Predict block + block from workspace.I have a learned CNN machine and I have data for prediction in the form of timetables and timeseries. I have 77 predictors and 9 sequences to test. Everything compiles and the graph/scope is strange T = 22, where these values should be T = 9. Some problem in my opinion with providing data. What should be the correct formula?
0 个评论
回答(1 个)
Ayush
2023-5-31
To provide data to a CNN machine in SIMULINK with a timestamp, the following steps can be followed:
1. Load the data into the workspace in MATLAB. Ensure that the data is in the form of either a timetable or timeseries.
2. Create a Predict block in the Simulink model. In the Predict block, select the trained CNN model that you want to use for prediction. You can read more about training the CNN here Train Convolutional Neural Network for Regression - MATLAB & Simulink (mathworks.com)
3. Add a "From Workspace" block to the Simulink model. In the block parameters, specify the data that you want to use for prediction. Make sure that the data is in the correct format and includes the timestamp information.
4. Connect the "From Workspace" block to the input of the Predict block.
5. Run the simulation and observe the results using the Scope or Graph block.
In terms of the strange results that you are observing in the Scope or Graph block, it is likely that there is an issue with the data format or input configuration. Check that the data is formatted correctly and that the input to the Predict block is configured properly. Also, ensure that the CNN model was trained using data that is similar in format and structure to the data being inputted for prediction.
Without seeing the exact configuration of your Simulink model and data, it is difficult to provide a specific formula for correcting the issue. However, carefully reviewing the data format and input configuration should help to identify and resolve the problem.
0 个评论
另请参阅
类别
在 Help Center 和 File Exchange 中查找有关 Sequence and Numeric Feature Data Workflows 的更多信息
产品
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!