I have a set of devices, each with a list of T times that are displayed when the device turns on, for example.
Device A: [Mon 16:03, Mon 15:59, Wed. 16:05, ... n]
I find patterns of use, for example, the next day, when a person turns on the switch on average T +/- 5 minutes, there will probably be a strong connection between this time and the average value of T. We can say that there is a pattern, and it can be created as the days go on. If there is a day without a value (the switch did not turn on), i.e. miss, then confidence can be diminished. One of the problems is that it will be necessary to take into account data that is missing for several days. We can say that if trust falls below the threshold, then the pattern does not exist.
I created a simple working version (without taking into account the mistakes), but I'm more interested in what a larger mind will consider as the best way to evaluate and determine if a daily event exists. I thought this was the best place for this, as I am interested in an elegant and beautiful way to approach this. Are there better statistical models that exist for developing such models? Thanks you
, ( fft) . 1 , , - ; 7 , , - , .
, . , - , , - ? , - (), (, ).
, , . google, , " " .
ps , . -, . , , , "5 " .. , , , , . - 5 , ... , , , , , .
pps . /, , .
, , e 10 24 , . . Boolean , e .
e . , - .