Prediction/Simulation with Matlab
2 ビュー (過去 30 日間)
I'm running some experiments in a lab on a machine and i'm trying to simulate the output based on these experiments and a governing equation that relates the materials used in the machine to the relative humidity in the atmosphere. I'm running the machine at different levels of relative humidity and recording the outputs for each individual material. I repeat the tests and get the output data with 300 trials for three different kinds of materials( 100 for each material, i.e, output for the individual material at five levels of relative humidity). I want to use this data along with the above mentioned governing equation to predict the output for a fourth material at differrent levels of humidity. Is this possible with the deep learning toolbox? or System identification or regression? Can i have some tips on where to start with this idea? I can look up the documentation and proceed but since i'm totally new to neural networks, it'd be helpful if someone could share some valuable information/tips on where to begin or the add-ons that i require for this.
https://docdro.id/eW0A8RH (the link has tables with only 15 trials but its supposed to be 300 trials (100 for each material)
回答 (1 件)
Iuliu Ardelean 2021 年 1 月 12 日
編集済み: Iuliu Ardelean 2021 年 1 月 12 日
A good place to start would be polyfit and polyval: https://uk.mathworks.com/help/matlab/ref/polyfit.html . This should be a good starting point for most applications.
Then, if you need more advanced models or a more thorough workflow, you can check out this post: https://uk.mathworks.com/help/stats/supervised-learning-machine-learning-workflow-and-algorithms.html.