I am new two time series related problem.
I have some time series data from imu sensors.
The time is given in seconds. I dnt have the date month year information.
For each time there is a label of activity like what happened or not.
Let say we have some data from sensor X. It would look the following example.
time, sensor_id, field 1, field2, field3 , … , Event
3.009, uixcd994, 0, 10, 4.5, …, Exercisee
and after hundreds for rows
3000.845, uixcd994, 0, 10, 4.5, …, Training
and after thousands of more rows some other event.
Remember that, Order can be anything,
Training → Exercise → Training or
Exercise → Training → Something Else → No_acitivity → Training.
However, we have maximum four possibility. Training, Exercise, Warm_up or No activity.
I have similar data from dozens of sensors. What I want to do is: For a given test data , I want to identify which part of the sensor data is what e,g, training exercise, etc.
However, I could not find a good example how to start and how to process the data .Since the problem sounds like both classification and regression. Any ideas, example ( for classification, regression or for both) would be really appreciated.