I have a dataset of EEG with 2 classes , normal subject and patients. Each subject has 3 conditions. It mean during data collection of each subject, the subject undergoes three conditions, You can say these three conditions are
- Standing
- Sitting
- Lying
So normal subject # 1 EEG data is recorded and he under goes these three conditions
I have converted this data into epoch and find their features
epoch 1- class 0, condition 1, subject 1
Epoch 2- class 0, condition 1, subject 2
and so on
If protocol has no conditions I can create table like this to feed into some machine learning algorithm
Subject Feature 1 Feature 2 Feature 3 Class
Subject 1 Value Value Value 0
Subject 2 Value Value Value 1
Subject 3 Value Value Value 0
Subject 4 Value Value Value 1
I am confused how to create data frame for machine learning algorithm when we have conditions
Do we just put the values of condition 1 and below it values of condition 2 and below it values of condition 3
For your example, you can use an array of three columns and three rows. the columns represent the subjects and the rows comprise the conditions.
no, the features, such as time domain features, mean, median, skewness, entropies etc
I think that each feature may be processed individually, or you may fined a relationship between these features. For example, the Probability Density Function (PDF) can be used to combine the mean and covariance features.
- Badges
- Science topic