Training the model with experimental data

Question

I want to train my SVR (support vector regression) model using experimental data. The experimental data is as follows:-

1. Each experiment results in a similar to sinusoidal wave output with X axis being time, sampling interval being 0.1 sec and total 200 sec so each experiment gives 2000 points.

2. If I repeat the experiment 100 times I will have 100 similar results with the same sampling intervals and 2000 points.

Now my problem is that I will have around 100 CSV files . How will I train my SVR model using these CSV files since whatever code I have seen they usually input 1 CSV file per fitting.

Also what would be the test train data split procedure for this kind of problem.

Answer 1

Import multiple csv files into pandas and concatenate into one DataFrame

You could use this link to read multiple csv files at once in a folder and combine them into 1 csv file assuming you have same columns in all those csv files. While concatenating different csv files, if you wish to have a unique code for each csv file then you could add the file name as one of the column in each CSV file. Later you can use one-hot encoding technique to make it a distinctive feature in the data. Also, while doing test train split you could assign 80 sheets as train and 20 sheets as test data.

Hope this helps!