1
$\begingroup$

I have a classification problem where I am planning to use hourly traffic data for a day. Is there any way to compress it? instead of creating 24 predictors which account for hourly traffic?

$\endgroup$
2
  • $\begingroup$ usual dimensionality reduction methods should go. it also depends on what is your task (what do you want to classify, how do you think it relates to traffic). you may also decide to sum all those columns together and only keep the sums column. $\endgroup$
    – carlo
    Nov 9, 2019 at 11:30
  • $\begingroup$ I have hourly traffic data of which I can use hours value. However, I am hoping to find the correlations between in flow and out flow. For example, a heavy traffic in two hours followed by none might indicate lots of population moving out of the region. I am hoping to find such patterns, can CNNs help to capture that info? As they are used for sensor data $\endgroup$
    – Shanthan K
    Nov 9, 2019 at 20:32

1 Answer 1

1
$\begingroup$

your problem is a panel data .since you have both time and cross-section. and yes panel data methods can help you have one predictor which considers time itself. (it is kind of like considering a 23dummy variables for each time, then applying them in one equation) you can read about (panel data or longitudinal data)and choose which methods can help you. good luck

$\endgroup$
3
  • $\begingroup$ I don't think this is really the same as usual panel data. each hour data is meaningful in respect to the hour it has been recorded in $\endgroup$
    – carlo
    Nov 9, 2019 at 11:27
  • $\begingroup$ i still think the data is panel. but maybe for this case it should be seprated to clusters according to times for example highly traffic hours or normal traffic or .. something like that. i also suggest to search for what uber use to analyse his data.lots of information about it is available they can inspiring @carlo $\endgroup$
    – Soma
    Nov 9, 2019 at 11:48
  • $\begingroup$ No, it is not a panel. I have hourly traffic data of which I can use only hours value. However, I am hoping to find the correlations between in flow and out flow. For example, a heavy traffic in two hours followed by none might indicate lots of population moving out of the region. I am hoping to find such patterns. $\endgroup$
    – Shanthan K
    Nov 9, 2019 at 20:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.